AI Database

AI and Market Power

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
AI and Market Power AI exclusion and exploitation	Algorithmic Pricing - A Competition Law Perspective on Personalised Prices	In the age of big data, sellers can amass a considerable amount of up-to-date information about their customers that algorithms can then use in order to engage in personalised pricing practices at a large scale. From a competition law perspective, several questions arise in this context. The prospect of an increased personalisation of prices may lead to a gradual disappearance of uniform market prices, which have been at the centre of economics-based competition law for many years. This could call into question many of the general tools of competition analysis. Furthermore, the question poses itself to what extent algorithmic personalised pricing could constitute an anti-competitive practice that can and should be challenged under current competition laws. While some theories of harm, like discrimination and excessive prices, are possible candidates when analysing the applicability of well-known abuses to algorithmic personalised pricing, consumer trust in digitalisation and markets may warrant the elaboration of a theory of harm specifically tailored to algorithmic personalised pricing. In addition, the question poses itself whether the particular harm that algorithmic personalised pricing can inflict should be considered in more tailored regulation. While this contribution’s focus is on EU competition law, similar principles can guide the discussion in other jurisdictions.	2023	Viktoria HSE Robertson	Algorithmic Pricing - A Competition Law Perspective on Personalised Prices
AI and Market Power AI exclusion and exploitation	Rethinking Algorithmic Explainability Through the Lenses of Intellectual Property and Competition	Algorithmic decision-making is integral to digital platforms, influencing user experiences and societal dynamics. This paper chapter scrutinizes algorithmic opacity, highlighting the inherent biases, the anti-competitive strategies that may result from dominant market power and the potential for discrimination within these systems. Despite the promise of objectivity, algorithms often operate under a veil of opacity, shaping content and information access, with significant implications for individual perspectives and societal functioning. The paper explores the legal challenges posed by the protection of algorithms through the lenses of intellectual property rights and competition law. It calls for a multifaceted regulatory approach to ensure transparency. The analysis emphasizes the need to balance innovation with competition and societal well-being, advocating for a right to explanation in the face of automated decisions within the European Union.	2024	Lucas Anjos	Rethinking Algorithmic Explainability Through the Lenses of Intellectual Property and Competition
AI facilitated manipulation AI and Market Power	Is Generative Ai the Algorithmic Consumer We are Waiting for?	Most studies on the competitive effects of Generative AI focus on the supply side. Interest in consumers is generally restricted to their roles as users of the technology, as well as indirect trainers of Generative AI models through their prompt engineering. In this short article, we focus instead on the potential effects of Generative AI on competition that arise from an active use of Generative AI on the demand side, by consumers seeking goods and services. In particular, we explore the possibility that Generative AI Large Language Models (LLMs) can act as truncated algorithmic consumers, assisting consumers in deciding which products and services to purchase, thereby potentially reducing consumers' information costs and increasing competition. We explore how LLMs' unique characteristics – mainly their conversational use and the provision of an authoritative single answer, as well as spillover trust effects from their other uses – might motivate consumers to use them to search for products and services. We then analyze some of the limitations and competition concerns that might result from the use of Generative AI by consumers. In particular, we show how LLMs' modus operandi - trained to seek the most plausible next word - lead to outcome homogenization and increase entry barriers for new competitors in product markets. We also explore the potential of manipulation and gaming of LLM models. As elaborated, a combination of an LLM model with a dataset on consumers' digital profiles might potentially create a strong nudging mechanism, recreating consumer choice architecture to optimize commercial goals and exploiting consumer’s behavioral biases in novel ways not envisioned.	2024	Michal Gal, Amit Zac	Is Generative Ai the Algorithmic Consumer We are Waiting for?
AI and Market Power AI exclusion and exploitation	Competitive Model Selection in Algorithmic Targeting	We consider competition between firms that design and use algorithms to target consumers. Firms first choose the design of a supervised learning algorithm in terms of the complexity of the model or the number of variables to accommodate. Each firm then appoints a data analyst to estimate demand for multiple consumer segments by running the chosen design of the algorithm. Based on the estimates, each firm devises a targeting policy to maximize estimated profit. The firms face the general trade-off between bias and variance in model selection. We show that competition may induce firms to choose algorithms with more bias leading to simpler (less flexible) algorithmic choice. This implies that complex (more flexible) algorithms such as deep learning that show greater variance in the estimates are more valuable to firms with greater monopoly power.	2023	Ganesh Iyer, T Tony Ke	Competitive Model Selection in Algorithmic Targeting
AI and Market Power	Algorithmic Product Positioning and Pricing: Can Artificial Intelligence Do Strategy?	Companies are increasingly interested in using Artificial Intelligence (AI) and algorithms, and there is substantial economic literature on algorithmic pricing, but can AI algorithms do well beyond pricing? We study the feasibility of using AI algorithms to determine product positioning and pricing autonomously. The novelty of our approach is that AI algorithms learn to make both decisions sequentially in a theoretical competitive environment. We test our approach in two cases with opposing equilibrium characteristics and find that the simplest version of Q-learning, a reinforcement learning algorithm, can optimally position a product and set prices accordingly. This adds to prior work that evaluates the performance of Q-learning in pricing. Our results demonstrate the plausibility of AI-enabled systems that make strategic decisions within companies. We discuss implications about the growing strategic role of AI within companies and identify the need for more research on algorithmic strategic management.	2024	J. Manuel Sánchez-Cartas, Evangelos Katsamakas	Algorithmic Product Positioning and Pricing: Can Artificial Intelligence Do Strategy?
AI and Market Power	The Strengthening of the Oligopoly Problem by Algorithmic Pricing	The purpose of the paper is to establish if algorithms have an anticompetitive impact on firms’ pricing behavior that may not be detected by competition law. Specifically, it examines how pricing algorithms change the structure on the market and how this strengthens the oligopoly problem.	2020	Amalie Toft Bentsen	The Strengthening of the Oligopoly Problem by Algorithmic Pricing
AI and Market Power	Data and Algorithms: Strategic Disclosure of Competitiveness on Platforms Through Marketplace Analytics	The data boom in e-commerce has spurred AI-powered marketplace analytics, but platforms hold the data reins. Some adopt open data-access policies with third-party analytics providers (e.g., permitting data-scraping or API-sharing) while others are restrictive. We ask why and when an e-commerce platform—capable of designing its own analytics to control sellers’ actions—may benefit from open data-access policies to accommodate competing third-party analytics services, despite the potential drawbacks of weakening its data advantages and control. We analyze two intertwined decisions an e-commerce platform can make when designing analytics to predict market competitiveness and assist sellers’ pricing decisions involving (1) data-access policy and (2) algorithm design. We find that platforms may use over-optimistic algorithms (by increasing the likelihood of generating low-competition signals) in their own analytics to boost commissions. Sellers, however, prefer a more accurate algorithm to confidently set higher prices based on reliable signals. This misaligned incentive may make sellers reluctant to adopt a platform’s analytics, resulting in a lose-lose situation and prompting the platform to allow data access to third-party providers. Overall, platforms gain from open data-access strategies in markets with moderately strong or weak competition. Finally, privacy legislation aimed at curtailing platforms’ data-sharing practices may inadvertently hurt consumers.	2025	Yi Liu, Fei Long	Data and Algorithms: Strategic Disclosure of Competitiveness on Platforms Through Marketplace Analytics
AI and Market Power	Competitive Algorithmic Targeting and Model Selection	We consider competition between firms that design and use algorithms to target consumers. Firms first choose the design of a supervised learning algorithm in terms of the complexity of the model or the number of variables to accommodate. Each firm then appoints a data analyst to estimate demand for multiple consumer segments by running the chosen design of the algorithm. Based on the estimates, each firm devises a targeting policy to maximize estimated profit. The firms face the general trade-off between bias and variance in model selection. We show that competition may induce firms to choose algorithms with more bias leading to simpler (less flexible) algorithmic choice. This implies that complex (more flexible) algorithms such as deep learning that show greater variance in the estimates are more valuable to firms with greater monopoly power.	2022	Ganesh Iyer, T. Tony Ke	Competitive Algorithmic Targeting and Model Selection
AI and Market Power AI exclusion and exploitation	How the Uber & Lyft Case Provides an Impetus to Re-Examine Buyer Power in the World of Big Data and Algorithms	For competition lawyers, Uber is an interesting subject to study. Not only does Uber change the dynamics of the transportation market but it also raises interesting competition law questions. Last year for example, a class action suit against Uber in New York raised the question whether Uber is possibly arranging a hub and spoke cartel amongst the drivers by coordinating their selling prices. 2017 has continued to be litigious and interesting. One of these new class action lawsuits might also raise thought-provoking antitrust issues related to big data and buyer power. Uber, the maverick firm that revolutionized passenger transportation services across the world has been now sued over its alleged use of its “Hell” software before the U.S. District Court for the Northern District of California filed on April 24th, 2017. The suit alleges a breach of privacy laws due to interception of private communications and unfair competition. This software apparently allowed Uber to track Lyft drivers, its main competitor, create fake Lyft accounts, determine which drivers drove for both companies, and “execut[e] a plan meant to entice double-appers to drive exclusively for them”. In this paper we explore such behaviour from a different perspective, the antitrust one. The focus of this paper is on exploring relevant behavior from a buyer power-oriented focusing on reverse rebates and overbuying, while not engaging in a concrete analysis of Uber’s conduct. This analysis provides us with the opportunity to re-explore traditional antitrust concepts, anchored on the purchasing of raw material, in the data and algorithm driven world, in particular, how companies can use big data in anticompetitive strategies, such as granting supra-competitive bonuses, overbuying, and raising rival’s costs through overbuying input.	2017	Ignacio Herrera Anchustegui, Julian Nowag	How the Uber & Lyft Case Provides an Impetus to Re-Examine Buyer Power in the World of Big Data and Algorithms
AI and Market Power AI exclusion and exploitation	Platform Preferencing and Price Competition II: Evidence From Amazon	A previous study of platform preferencing is extended to an environment in which firm prices are determined in a non-myopic, dynamic setting, which depends on the platform's preferencing rule. These dynamic pricing strategies result in pricing patterns that closely resemble Edgeworth cycles. I provide a method to estimate the primitives that govern these pricing cycles, which allows me to assess the counterfactual welfare implications associated with various preferencing rules, using a large dataset from the Amazon platform from 2018-2022. The welfare effects of policy proposals in this environment, in particular those that eliminate self-preferencing, significantly differ from those under the static price competition environment.	2025	Olivia Hartzell	Platform Preferencing and Price Competition II: Evidence From Amazon
AI and Market Power AI exclusion and exploitation AI facilitated collusion	Virtual Competition: The Promise and Perils of the AlgorithmDriven Economy	Shoppers with Internet access and a bargain-hunting impulse can find a universe of products at their fingertips. In this thought-provoking exposé, Ariel Ezrachi and Maurice Stucke invite us to take a harder look at today’s app-assisted paradise of digital shopping. While consumers reap many benefits from online purchasing, the sophisticated algorithms and data-crunching that make browsing so convenient are also changing the nature of market competition, and not always for the better. Computers colluding is one danger. Although long-standing laws prevent companies from fixing prices, data-driven algorithms can now quickly monitor competitors’ prices and adjust their own prices accordingly. So what is seemingly beneficial―increased price transparency―ironically can end up harming consumers. A second danger is behavioral discrimination. Here, companies track and profile consumers to get them to buy goods at the highest price they are willing to pay. The rise of super-platforms and their “frenemy” relationship with independent app developers raises a third danger. By controlling key platforms (such as the operating system of smartphones), data-driven monopolies dictate the flow of personal data and determine who gets to exploit potential buyers. Virtual Competition raises timely questions. To what extent does the “invisible hand” still hold sway? In markets continually manipulated by bots and algorithms, is competitive pricing an illusion? Can our current laws protect consumers? The changing market reality is already shifting power into the hands of the few. Ezrachi and Stucke explore the resulting risks to competition, our democratic ideals, and our economic and overall well-being.	2016	Ariel Ezrachi, Maurice E Stucke	Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy
AI and Market Power	The Power of the bargaining Robot	Antitrust is worried about the potential of recent advances in technology to increase market power, which is the ability of a firm to undermine competition from sellers of competing brands. Recent advances are indeed creating opportunities for firms to enhance their market power. But as the cost of robots falls to rates affordable by small firms, technology also promises to eliminate many economies of scale, reducing market power. The net effect of technological advance on market power is therefore not determinate in the long run	2017	Ramsi A Woodcock	The Power of the bargaining Robot
AI and Market Power AI mergers and coopertation AI exclusion and exploitation AI facilitated collusion	Big Data and Competition Policy	Big Data and Big Analytics are a big deal today. Big Data is playing a pivotal role in many companies' strategic decision-making. Companies are striving to acquire a 'data advantage' over rivals. Data-driven mergers are increasing. These data-driven business strategies and mergers raise significant implications for privacy, consumer protection and competition law. At the same time, European and United States' competition authorities are beginning to consider the implications of a data-driven economy on competition policy. In 2015, the European Commission launched a competition inquiry into the e-commerce sector and issued a statement of objections in its Google investigation. The implications of Big Data on competition policy will likely be a part of the mix. Big Data and Competition Policy is the first work to offer a detailed description of the important new issue of Big Data and explains how it relates to competition laws and policy, both in the EU and US. . The book helps bring the reader quickly up to speed on what is Big Data, its competitive implications, the competition authorities' approach to data-driven mergers and business strategies, and their current approach's strengths and weaknesses. Written by two recognized leading experts in competition law, this accessible work offers practical guidance and theoretical discussion of the potential benefits (including data-driven efficiencies) and concerns for the practitioner, policy maker, and academic alike.	2016	Maurice Stucke, Allen Grunes	Big Data and Competition Policy
AI and Market Power	The Google Book Search Settlement. A New Orphan-Works Monopoly?	This paper considers the proposed settlement agreement between Google and the Authors Guild relating to Google Book Search. Google boldly launched Google Book Search in pursuing its goal of organizing the world’s information. Even though Google was sensitive to copyright values, the service relied on mass copying and thus Google undertook a substantial legal risk in setting up the service. That risk was realized with the lawsuits by the Authors Guild and the Association of American Publishers. The October, 2008 settlement agreement for those suits will create an important new copyright collective and will legitimate broad-scale online access to United States books registered before early January, 2009. The settlement agreement is exceeding complex but I have focused on three issues that raise antitrust and competition policy concerns. First, the agreement calls for Google to act as agent for rights holders in setting the price of online access to consumers. Google is tasked with developing a pricing algorithm that will maximize revenues for each of those works. Direct competition among rights holders would push prices towards some measure of costs and would not be designed to maximize revenues. As I think that that level of direct coordination of prices is unlikely to mimic what would result in competition, I have real doubts about whether the consumer access pricing provision would survive a challenge under Section 1 of the Sherman Act. Second, and much more centrally to the settlement agreement, the opt out class action will make it possible for Google to include orphan works in its book search service. Orphan works are works as to which the rightsholder can’t be identified or found. That means that a firm like Google can’t contract with an orphan holder directly to include his or her work in the service and that would result in large numbers of missing works. The opt out mechanism - which shifts the default from copyright’s usual out to the class action’s in - brings these works into the settlement. But the settlement agreement also creates market power through this mechanism. Absent the lawsuit and the settlement, active rights holders could contract directly with Google, but it is hard to get large-scale contracting to take place and there is, again, no way to contract with orphan holders. The opt out class action then is the vehicle for large-scale collective action by active rights holders. Active rights holders have little incentive to compete with themselves by granting multiple licenses of their works or of the orphan works. Plus under the terms of the settlement agreement, active rights holders benefit directly from the revenues attributable to orphan works used in GBS. We can mitigate the market power that will otherwise arise through the settlement by expanding the number of rights licenses available under the settlement agreement. Qualified firms should have the power to embrace the going-forward provisions of the settlement agreement. We typically find it hard to control prices directly and instead look to foster competition to control prices. Non-profits are unlikely to match up well with the overall terms of the settlement agreement, which is a share-the-revenues deal. But we should take the additional step of unbundling the orphan works deal from the overall settlement agreement and create a separate license to use those works. All of that will undoubtedly add more complexity to what is already a large piece of work, and it may make sense to push out the new licenses to the future. That would mean ensuring now that the court retains jurisdiction to do that and/or giving the new Registry created in the settlement the power to do this sort of licensing. Third, there is a risk that approval by the court of the settlement could cause antitrust immunities to attach to the arrangements created by the settlement agreement. As it is highly unlikely that the fairness hearing will undertake a meaningful antitrust analysis of those arrangements, if the district court approves the settlement, the court should include a clause - call this a no Noerr clause - in the order approving the settlement providing that no antitrust immunities attach from the court’s approval.	2009	Randal C Picker	The Google Book Search Settlement. A New Orphan-Works Monopoly?
AI and Market Power	Should We Be Concerned That Data And Algorithms Will Soften Competition?	-	2017	Paul A Johnson	Should We Be Concerned That Data And Algorithms Will Soften Competition?
AI and Market Power AI mergers and cooperation	EU Competition Law in the Sharing Economy	This article provides an analytical framework for the competition law assessment of activities in the sharing economy. It is argued that sharing economy platforms are two-sided businesses active in intermediation. Sharing economy intermediation markets are likely to become concentrated and possibly dominated by a single market player. The activities of powerful sharing economy platforms, for which data use is key, are likely to be scrutinised in merger control proceedings and in the long term potentially also in the area of market abuse. This is an ongoing competition law analysis and needs to be re-evaluated in light of constantly developing market circumstances.	2016	Guy Lougher , Sammy Kalmanowicz	EU Competition Law in the Sharing Economy
AI and Market Power AI mergers and cooperation	Google: The Unique Case of the Monopolistic Search Engine	-		Albert A Foer and Sandeep Vaheesan	Google: The Unique Case of the Monopolistic Search Engine
AI and Market Power	Is Big Data a Different Kind of Animal? The Treatment of Big Data Under the EU Competition Rules	More frequently than not, competition law is shaped by trends. ‘Big Data’ has been one of the hottest trending topics in the competition world during the past few years, monopolising the theme of numerous conferences, speeches, and academic articles. Unsurprisingly, therefore, Big Data has also been increasingly at the forefront of the Commission's and national competition authorities’ enforcement priorities. For instance, the Commission is currently assessing whether existing competition enforcement tools are sufficient and adequate to address possible competition concerns arising from companies’ accumulation of ‘unique’, strategic user data that could enhance market power, create barriers to entry and, ultimately, foreclose rivals. Moreover, the Commission, as well as national competition authorities, have voiced their intention to be vigilant regarding possible anticompetitive use of such data by dominant firms and increased market transparency that could facilitate collusion between rivals. The purpose of this article is to question and critically assess whether the issues that may arise in relation to Big Data are truly unique and pressing, such that cases involving Big Data should be treated differently under competition rules or even regulation or whether the existing tools are sufficient and adequate to address any such issues.	2017	Marixenia Davilla	Is Big Data a Different Kind of Animal? The Treatment of Big Data Under the EU Competition Rules
AI and Market Power	Competition Law and Data	The collection, processing and commercial use of data is often seen not as a competition law issue but rather as an issue which concerns data protection enforcement. However, several recent proceedings point to the fact that competition authorities have begun to look at possible competition issues arising from the possession and use of data, even if, in the end, none were ascertained in the specific cases. Recent developments in digital markets have led to the emergence of a number of firms that achieve extremely significant turnovers based on business models which involve the collection and commercial use of (often personal) data. Some of them enjoy a very high share of users in the service sector in which they are active. The Google search engine and the Facebook social network are probably the most prominent examples. While many of the services provided by these firms are marketed as ‘free’, their use involves in practice making possible the collection of personal information about the users. This has spurred new discussions about the role of data in economic relationships as well as in the application of competition law to such relationships, in particular as regards the assessment of data as a factor to establish market power. It is important to note that although these questions are often examined with the examples of Google and of Facebook in mind, they are also relevant for many other industries. Indeed, the development of data collection already goes well beyond search engines, social networking or online advertising and extends today to sectors such as energy, telecommunications, insurance, banking or transport. Furthermore, in the near future, the development of connected devices should make data more and more relevant for product industries and not only for services. This paper aims to feed this debate by identifying some of the key issues and parameters that may need to be considered when assessing the interplay between data, market power and competition law. For this purpose it is necessary to first clarify what can be meant by “data” or the often cited “big data”, whether there are different types of data with possibly different features, in which possible ways data can be collected and how they are used by firms (section II). The various theories of harm usually associated with data collection and exploitation in digital markets are presented in section III. Finally, in view of these two sections, section IV discusses some of the parameters that are to be considered in assessing the relevance and credibility of these theories of harm.	2016	Autorité de la concurrence, Bundeskartellamt	Competition Law and Data
AI and Market Power	AI Pricing Algorithms Under Platform Competition	Platforms play an essential role in the modern economy. At the same time, due to advances in artificial intelligence (AI), algorithms are becoming more widely used for pricing and other business functions. Previous literature examined algorithmic pricing, but not in the context of network effects and platforms. Moreover, platform competition literature has not considered how algorithms may affect competition. We study the performance of AI pricing algorithms (Q-learning and Particle Swarm Optimization) and naïve algorithms (price-matching) under platform competition. We find that algorithms set an optimal price structure that internalizes network effects. However, no algorithm is always the best because profitability depends on the type of competing algorithms and market characteristics, such as differentiation and network effects. Additionally, algorithms learn autonomously when an equilibrium is unstable and avoid it. When algorithm adoption is an endogenous strategic decision, several algorithms can be adopted in equilibrium; we characterize the conditions for the various outcomes and show that the equilibrium and platform profits are sensitive to algorithm design changes. Overall, our research suggests that AI algorithms can be effective in the presence of network effects, and platforms are likely to adopt a variety of algorithms. Lastly, we reflect on the business value of AI and identify opportunities for future research at the intersection of AI algorithms and platforms.	2023	J. Manuel Sánchez-Cartas, Evangelos Katsamakas	AI Pricing Algorithms Under Platform Competition
AI and Market Power AI exclusion and exploitation	What Should Be Done About Google's Quasi-Monopoly in Search? Mandatory Data Sharing Versus AI-Driven Technological Competition	The first part of this paper focuses on competition between search engines that match user queries with webpages. User welfare, as measured by click-through rates on top-ranked pages, increases when network effects attract more users and generate economies of scale in data aggregation. However, network effects trigger welfare concerns when a search engine reaches a dominant market position. The EU Digital Markets Act (DMA) imposes asymmetric data sharing obligations on very large search engines to facilitate competition from smaller competitors. We conclude from the available empirical literature on search-engine efficiency that asymmetric data sharing may increase competition but may also reduce scale and user welfare, depending on the slope of the search-data learning curve. We propose policy recommendations to reduce tension between competition and welfare, including (a) symmetric data sharing between all search engines irrespective of size, and (b) facilitate user real-time search history and profile-data portability to competing search engines. The second part of the paper focuses on the impact of recent generative AI models, such as Large Language Models (LLMs), chatbots and answer engines, on competition in search markets. LLMs are pre-trained on very large text datasets, prior to usage. They do not depend on user-driven network effects. That avoids winner-takes-all markets. However, high fixed algorithmic learning costs and input markets bottlenecks (webpage indexes, copyright-protected data and hyperscale cloud infrastructure) make entry more difficult. LLMs produce semantic responses (rather than web pages) in response to a query. That reduces cognitive processing costs for users but may also increase ex- post uncertainty about the quality of the output. User responses to this trade-off will determine the degree of substitution or complementarity between search and chatbots. We conclude that, under certain conditions, a competitive chatbot markets could crowd out a monopolistic search engine market and may make DMA-style regulatory intervention in search engines redundant. The paper concludes with some policy recommendations.	2023	Bertin Martens	What Should Be Done About Google's Quasi-Monopoly in Search? Mandatory Data Sharing Versus AI-Driven Technological Competition
AI and Market Power AI mergers and cooperation	Competition in Pricing Algorithms	We document new facts about pricing technology using high-frequency data, and we examine the implications for competition. Some online retailers employ technology that allows for more frequent price changes and automated responses to price changes by rivals. Motivated by these facts, we consider a model in which firms can differ in pricing frequency and choose pricing algorithms that are a function of rivals' prices. In competitive (Markov perfect) equilibrium, the introduction of simple pricing algorithms can generate price dispersion, increase price levels, and exacerbate the price effects of mergers.	2021	Zach Brown, Alexander MacKay	Competition in Pricing Algorithms
AI and Market Power AI facilitated collusion	Algorithmic Pricing and Liquidity in Securities Markets	We study ``Algorithmic Market Makers" (AMs) that use Q-learning algorithms to set prices for a risky asset. We find that while AMs successfully adapt to adverse selection, they struggle to learn competitive pricing strategies. This failure is driven by limited experimentation and noisy feedback regarding the profitability of undercutting a competitor. Consequently, an increase in AMs' profit volatility tends to result in less competitive market outcomes. These features leave identifiable patterns: for example, AMs earn higher rents in the absence of adverse selection, and their bid-ask spreads respond asymmetrically to symmetric shocks to their costs.	2025	Stefano Lovo, Jean-Edouard Colliard, Thierry Foucault	Algorithmic Pricing and Liquidity in Securities Markets
AI and Market Power AI exclusion and exploitation	Machine Learning as Natural Monopoly	Machine learning is transforming the economy, reshaping operations in communications, law enforcement, and medicine, among other sectors. But all is not well: Many machine-learning-based applications harvest vast amounts of personal information and yield results that are systematically biased. In response, policy makers have begun to offer a range of incomplete solutions. In so doing, they have overlooked the possibility—suggested intuitively by scholars across disciplines—that these systems are natural monopolies and have thus neglected the long legal tradition of natural monopoly regulation. Drawing on the computer science, economics, and legal literatures, I find that some machine-learning-based applications may be natural monopolies, particularly where the fixed costs of developing these applications and the computational costs of optimizing these systems are especially high, and where network effects are especially strong. This conclusion yields concrete policy implications: Where natural monopolies exist, public oversight and regulation are typically superior to market discipline through competition. Hence, where machine-learning-based applications are natural monopolies, this regulatory tradition offers one framework for confronting a range of issues—from privacy to accuracy and bias—that attend to such systems. Just as prior natural monopolies—the railways, electric grids, and telephone networks—faced rate and service regulation to protect against extractive, anticompetitive, and undemocratic behaviors, so too might machine-learning-based applications face similar public regulation to limit intrusive data collection and protect against algorithmic redlining, among other harms.	2022	Tejas N Narechania	Machine Learning as Natural Monopoly
AI and Market Power AI exclusion and exploitation	AI Adoption in a Monopoly Market	The adoption of artificial intelligence (AI) prediction of demand by a monopolist firm is examined. It is shown that, in the absence of AI prediction, firms face complex trade-offs in setting price and quantity ahead of demand that impact on the returns of AI adoption. Different industrial environments with differing flexibility of prices and/or quantity ex post, also impact on AI returns as does the time horizon of AI prediction. While AI has positive benefits for firms in terms of profitability, its impact on average price and quantity, as well as consumer welfare, is more nuanced and critically dependent on environmental characteristics.	2022	Joshua S Gans	AI Adoption in a Monopoly Market
AI and Market Power AI exclusion and exploitation	Antitrust & AI supply chains	Will AI technology disrupt the current Big Tech Barons, foster competition, and ensure future disruptive innovation that improves our wellbeing? Or might the technology help a few ecosystems become even more powerful? The article outlines the current digital market dynamics that lead to winner-take-most-or-all ecosystems. After examining the emerging AI foundation model supply chain, we consider several potential antitrust risks that may arise should specific layers become concentrated and firms extend their power across layers. After raising several countervailing factors that might lessen or prevent these antitrust risks, it conclude with suggestions for the policy agenda to promote both healthy competition and innovation in the AI supply chain.	2025	Maurice E. Stucke and Ariel Ezrachi	Antitrust & AI supply chains

AI facilitated exclusion and exploitation

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
AI facilitated exclusion and exploitation AI and Market Power	Behavior-Based Algorithmic Pricing	This article studies the impact of algorithmic pricing on market competition when firms collect data to charge personalized prices to their past customers. Pricing algorithms offer to each firm a rich set of pricing strategies combining first and third-degree price discrimination: they can choose for each of their past customers whether to charge them personalized or homogeneous prices. The optimal targeting strategy of each firm consists in charging personalized prices to past customers with the highest willingness to pay and a homogeneous price to the remaining consumers, including past customers with a low valuation on whom a firm has information. This targeting strategy maximizes rent extraction while softening competition between firms compared to classical models where firms target all past customers. In turn, price-undercutting and poaching practices are not sustainable with behavior-based algorithmic pricing, resulting in greater industry profits.	2023	Antoine Dubus	Behavior-Based Algorithmic Pricing
AI exclusion and exploitation AI and Market Power	Algorithmic Personalized Pricing: A Personal Data Protection and Consumer Law Perspective	Price is often the single most important term in consumer transactions. As the personalization of e-commerce continues to intensify, the law and policy implications of algorithmic personalized pricing i.e., to set prices based on consumers’ personal data with the objective of getting as closely as possible to their maximum willingness to pay (APP), should be top of mind for regulators. This article looks at the legality of APP from a personal data protection law perspective, by first presenting the general legal framework applicable to this commercial practice under competition and consumer law. There is value in analysing the legality of APP through how these bodies of law interact with one and the other. This article questions the legality of APP under personal data protection law, by its inability to effectively meet the substantive requirements of valid consent and reasonable purpose. Findings of illegality of APP under personal data protection law may in turn further inform the lawfulness of APP under competition and consumer law.	2024	Pascale Chapdelaine	Algorithmic Personalized Pricing: A Personal Data Protection and Consumer Law Perspective
AI exclusion and exploitation AI and Market Power	The Effect of Outsourcing Pricing Algorithms on Market Competition	A third party developer designs and sells a pricing algorithm that enhances a firm's ability to tailor prices to a source of demand variation, whether high-frequency demand shocks or market segmentation. The equilibrium pricing algorithm is characterized that maximizes the third party's profit given firms' optimal adoption decisions. Outsourcing the pricing algorithm does not reduce competition but does make prices more sensitive to the demand variation, and this is shown to decrease consumer welfare and increase industry profit. This effect is larger when products are more substitutable.	2021	Joseph E. Harrington Jr	The Effect of Outsourcing Pricing Algorithms on Market Competition
AI exclusion and exploitation AI facilitated collusion	Protecting consumers from collusive prices due to AI	The efficacy of a market system is rooted in competition. In striving to attract customers, firms are led to charge lower prices and deliver better products and services. Nothing more fundamentally undermines this process than collusion, when firms agree not to compete with one another and consequently consumers are harmed by higher prices. Collusion is generally condemned by economists and policy-makers and is unlawful in almost all countries. But the increasing delegation of price-setting to algorithms (1) has the potential for opening a back door through which firms could collude lawfully (2). Such algorithmic collusion can occur when artificial intelligence (AI) algorithms learn to adopt collusive pricing rules without human intervention, oversight, or even knowledge. This possibility poses a challenge for policy. To meet this challenge, we propose a direction for policy change and call for computer scientists, economists, and legal scholars to act in concert to operationalize the proposed change.	2020	Emilio Calvano, Giacomo Calzolari , Vincenzo Denicolò, Joseph E. Harrington Jr., Sergio Pastorello	Protecting consumers from collusive prices due to AI
AI exclusion and exploitation	Consumer’s Search In The Era Of Big Data	Consumer’s choice requires the collection of information to make a conscious and satisfactory decision. This structural feature of consumption has dramatically changed with the Internet and the diffusion of big data. This note reviews the impact of web-based searches on consumers’ satisfaction and surplus, distinguishing the case of search and experience goods.	2017	Michele Polo	Consumer’s Search in the Era of Big Data’
AI exclusion and exploitation	When algorithms set prices: winners and losers	The digital revolution has led to a significant growth in companies’ ability to capture, store and analyse data about their customers, competitors and the wider world. Increasingly, companies are using this information to develop algorithms that set prices for them. But how might the automation of pricing through algorithms affect competitive outcomes in markets, and result in different consumers being charged different amounts for the same good or service?	2017	Oxera	When algorithms set prices: winners and losers
AI exclusion and exploitation	Big Data: Bringing Competition Policy to the Digital Era	Business models based on the vast collection and process of user data in nearly real-time in recent years have enabled companies to offer a wide range of innovative and customised services, often at zero prices, with substantial gains for consumers. At the same time, data-driven network effects reinforced by user feedback loops, and high economies of scale associated with information technology infrastructures, may provide companies that own the data with market power and create a tendency for markets to tip. Concern is rising that the increasing reliance and use of personal data is harmful to consumers. While some practitioners have proposed adapting competition tools and antitrust policy to tackle such issues, others believe that these can be better addressed by data and/or consumer protection agencies. This paper attempts to define Big Data and its role within a competition context, and then identifies some of the potential implications for the enforcement of competition law in the areas of merger review, abusive of dominance and cartels. It also discusses how regulations on data ownership, access and portability may affect consumer protection and competitive neutrality. It was prepared as background for an OECD discussion held in November 2016.	2016	OECD	Big Data: Bringing Competition Policy to the Digital Era
AI exclusion and exploitation	The rise of behavioural discrimination’	The ongoing developments in e-commerce, big data and big analytics have transformed our online environment and the way we shop for goods and services. By increasing transparency, access to markets, and by reducing market barriers and our search costs, technological developments promise to lower the prices we pay, increase the selection of goods and services we are offered, and yield greater innovation. Indeed, we all expect to be better off in comparison to past decades when competition was less intense and largely confined to local offering by brick-and-mortar shops. And yet, is it possible that the initial promise of online competitiveness may give way to new dynamics that reduce our welfare? Are we still the winners in this story of innovation, or have we become targets of a new form of discrimination that increasingly extracts our wealth? In the online world, our anonymity and ability to identify a single competitive price are becoming a thing of the past. Virtual competition heralds the age of personalisation with its benefits and possible pitfalls. As a White House report summarised "[s]ellers are now using big data and digital technology to explore consumer demand, to steer consumers towards particular products, to create targeted advertising and marketing offers, and in a more limited and experimental fashion, to set personalised prices." Our article explores how e-commerce and the personalisation of our online environment can give rise to behavioural discrimination, a durable, more pernicious form of price discrimination. Online behavioural discrimination, as we explore, will likely differ from the price discrimination we have seen in the brickand-mortar retail world in several important respects: First is the shift from third-degree, imperfect price discrimination to near perfect price discrimination; second is the overall increase in consumption as the demand curve shifts to the right; and third is the durability of behavioural discrimination.	2016	Ariel Ezrachi, Maurice E Stucke	The rise of behavioural discrimination
AI exclusion and exploitation AI and Market Power	Looking Up in the Data-Driven Economy	With the rise of the super-platforms, we tend to look down (on their effect on consumers) rather than up (their effect on sellers and upstream providers). In looking down it seems like Google, Amazon and Facebook are using their power in the marketplace to deliver great value to us — wrestling lower prices from producers in the case of Amazon, bringing news onto a single platform in the case of Facebook, and organizing the world’s information, in the case of Google. While these companies appear to be furthering our interests, a closer look reveals how these super-platforms may wield their power downstream to harm us, the consumer. As our book Virtual Competition explores, the super-platforms can use our personal data to better price discriminate and their disincentive to protect our privacy (and promote technologies that do). Less discussed, but of significant concern, are the upstream effects of these super-platforms. They have the power to harm many of the companies from whom they buy or acquire content — and that harm ultimately harms us. In looking up rather than down, we see how the super-platforms can squeeze millions of sellers, including photographers, photojournalists, writers, journalists and musicians. Our competition laws deal with this kind of buyer power. These concerns, however, are often low on the enforcement agenda due to the indirect effects on “consumer welfare,” which is often measured by the price you pay for the goods or service. So if we stream the YouTube video ostensibly for “free,” the assumption is that our welfare is maximized. In the digital age, as this essay argues, that urgently needs to change	2017	Ariel Ezrachi, Maurice E Stucke	Looking Up in the Data-Driven Economy
AI exclusion and exploitation AI and Market Power	Is Your Digital Assistant Devious	Who wouldn’t want a personal butler? Technological developments have moved us closer to that dream. The rise of digital personal assistants has already changed the way we shop, interact and surf the web. Technological developments and artificial intelligence are likely to further accelerate this trend. Indeed, all of the leading online platforms are currently investing in this technology. Apple’s Siri, Amazon’s Alexa, Facebook’s M, and Google Assistant can quickly provide us with information, if we so desire, and anticipate and fulfill certain needs and requests. Yet, could they also reduce our welfare? Could they limit competition and transfer our wealth to the providers? And if so, can competition law safeguard our welfare while enabling these technological developments?	2016	Ariel Ezrachi, Maurice E Stucke	Is Your Digital Assistant Devious
AI exclusion and exploitation AI and Market Power	How Digital Assistants Can Harm our Economy, Privacy, and Democracy	Digital assistants embody the dream of an effortless future, free from the shackles of yesteryear: a tool which caters to users’ needs, excels at anticipating their wants, and delivers a personalized online environment. While digital assistants can certainly offer great value, a closer look reveals how—in an algorithm and data–driven world—a dominant digital assistant may ultimately serve the interests of corporations rather than consumers. Such assistants may be used to establish a controlled and manipulated personalized environment in which competition, welfare, privacy, and democracy give way to corporate interests. The future is not necessarily bleak, but requires our attention if users want the leading assistants to match the effortless dream.	2017	Maurice E Stucke, Ariel Ezrachi	How Digital Assistants Can Harm our Economy, Privacy, and Democracy
AI exclusion and exploitation AI and Market Power	Data is Infrastructure	Data is a contextual phenomenon. It reflects the social and material context from which it is derived and in which it is generated. It embeds the purposes, assumptions and rationales of those who produce, collect, use, share and monetize it. In the AI and digital platform economy, data’s role is primarily infrastructural. Its core uses are internal to companies. Data only rarely serves as a medium of exchange or commodity, and more frequently serves to profile users, train models, produce predictions, bundle and extend product capabilities which in turn are sold to advertisers and other customers. Insofar as they focus on the former, many technical, economic and legal attempts at defining data have inspired reductive policy efforts that include data protection, data ownership and limited data sharing remedies. This paper argues that understanding data as part of infrastructural pipelines can have significant conceptual and policy implications, and can redirect the way privacy, property and antitrust experts understand and govern data. This argument becomes more salient as market actors and regulators grapple with the catalyzing effects of neural networks and generative AI models on digital markets. In antitrust and competition law especially, regulators are consciously adopting a view of data as an infrastructural input into AI and other digital markets. Treating data as an input over which certain firms have competitive advantages can have significant implications for nascent AI markets, and yet the views in antitrust remain too narrow. Understanding data infrastructurally means viewing it not only as a critical input but also as inseparable from other material digital resources such as protocols, algorithms, semiconductors, and platform interfaces; as having important collective functions; and as calling for public interest regulation. Understanding data as infrastructure can move us past limited legal efforts and remedial solutions such as data separations, data sharing, and individual controls, and help re-orient how data is produced, stored and managed toward public uses.		Elettra Bietti	Data is Infrastructure
AI exclusion and exploitation AI and Market Power	Data Protection and Competition Law: The Dawn of ‘Uberprotection’	Our contribution aims at giving a comprehensive overview over the intertwining of competition law and data protection law in the EU legal framework, prompted by the rising and disruptive importance of amassed data, including personal data (‘Big Data’), for competition. Big Data has quickly penetrated most business areas in the past decade, posing challenges for the effectiveness of existing data protection rules, on one hand, but also for different aspects of competition law and its enforcement, on the other hand. Access to customer contact data or customer preferences has impacted on competitive parameters, raising completely new questions of competition law, e.g. in the context of data portability or digital cartels. However, the more fundamental issue arises if and how data protection compliance can or should be a parameter in the assessment of competition authorities around the world, being a well-known fact that, in principle, competitive assessment is bound only by welfare considerations. Personal data has had multiple impacts on all pillars of competition law – anticompetitive agreements, abuse of dominance and merger control. While abuse of dominance and merger control relate to competitive harm via the access to greater customer data, the classic price fixing cartels are being replaced by seemingly irretraceable, big data based price fixing algorithms. Proceeding empirically, from the recent practice of several competition authorities, including the European Commission, as well as the German and French competition authorities, this contribution identifies three phases of development of the intersection between data protection and competition law. At the beginning, competition authorities were acknowledging data protection law as being a separate issue without relevance for the purpose of merger control proceedings and thus placing the two areas of law on parallel pathways. In a second phase, the realization that data protection rules may in fact have a role in hampering or enabling competition took more and more space both in policymaking and in adjudication, with Data Protection Authorities starting to play a role. We are currently at the dawn of the third phase identified: data protection law considerations are at the core of at least one current case looking into an abuse of dominant position in Germany, a Digital Clearinghouse started to meet twice every year in Brussels and the European Commission is laying the groundwork for potential changes in the way it enforces competition law in the digital economy. We call this last phase “Uberprotection”, which we define as the protection of the rights of individuals and their welfare as data subjects, participants to the market and consumers, afforded by concerted enforcement and facilitated by coherent policymaking of competition authorities, data protection authorities and consumer protection authorities.	2018	Gabriela Zanfir-Fortuna, Sinziana Ianc	Data Protection and Competition Law: The Dawn of ‘Uberprotection’
AI exclusion and exploitation AI and Market Power	Competition Law for the Digital Era: A Complex Systems’ Perspective	As the global economy incurs a process of transformation by the ongoing ‘fourth industrial revolution’, competition law is traversing a ‘liminal’ moment, a period of transition during which the normal limits to thought, self-understanding and behaviour are relaxed, opening the way to novelty and imagination, construction and destruction. There is need for the discussion over the role of competition law in the digital era to be integrated to the broader debate over the new processes of value generation and capture in the era of digital capitalism and the complex economy to which it has given rise to. This complex digital economy is formed by a spider web of economic links, but also their underpinning societal relations, between different agents. However, competition law still lives in the simple world of neo-classical price theory (NPT) economics, which may not provide adequate tools in order to fully comprehend the various dimensions of the competition game. The emphasis put recently by competition authorities on multi-sided markets in order to analyse restrictions of competition in the data economy illustrates the agents’ changing roles and the complexity of their interactions, as the same agents can be at the same time consumers and producers while their personal data raw material for the value generation process. It becomes therefore essential to uncover the new value capture and value generation processes in operation in the digital economy, and draw lessons for the optimal design and enforcement of competition law, rather than take the established competition law framework as a given and try to stretch within it a quite complex reality that may not fit this Procrustean iron bed. These approaches should engage with the complex economics of digital capitalism, and in particular the role of futurity and financialisation, personalisation and cybernetics. These new developments, first, call for a re-conceptualisation of the goals of competition law in the digital era, as competition law moves from the calm and predictable waters of ‘consumer welfare’, narrowly defined, to integrate considerations of income/wealth distribution, privacy and complex equality. Second, it also requires a revision of the current understanding of the nature of the competitive game, which only focuses on horizontal rivalry in product and eventually technology markets. This is of course an important dimension of competition, but hardly the most significant one in the current process of value generation and capture in the digital economy. Firms do not only compete on the product market dimension, but in the today’s financialised economy, probably the most important locus of competition is capital markets. The process of financialisation has important implications for the development of digital capitalism, an issue that the paper explores in detail for the first time in competition law and economics scholarship. Financial markets evaluate companies in view of expected returns in the not so near future, often linked to the emergence of bottlenecks or the perception that a firm holds important assets and resources (e.g. data, algorithms, specialised labour). The role of financial markets’ evaluation in driving business strategies in the era of digital and financialised capitalism is linked to the ‘subtle shift of mindset’ in digital capitalism ‘from profit (and isolating mechanisms) to wealth creation (and the potential for asset appreciation)’ as value is created by investing in assets that will appreciate. Third, this calls for a consideration, not only of horizontal competition, but also of vertical competition, the competition for a higher percentage of the surplus value brought by innovation, and competition from complementary technologies that may challenge the lead position in the value chain of the incumbents (vertical innovation competition). Fairness considerations, among other reasons, may also lead competition authorities to not only focus on inter-platform/ecosystem competition but to also promote intra-platform/ecosystem competition, as this may be a significant element of the competitive game. To implement this broader focus of competition law, we need to develop adequate conceptual tools and methodologies. A recurrent problem is the narrow definition of market power in competition law, whose presence often triggers the competition law assessment, and which is also intrinsically linked to the step of market definition. This currently ignores possible restrictions of vertical competition, personalisation and the predictive role of digital platforms, which may become source of harm for consumers, the competitive process, or the public at large. It is important to engage with concepts of vertical power and the paper develops a typology of vertical power, combining in an overall conceptual framework the various concepts of non-structural power that have been used so far in competition law literature and some new ones (positional and architectural power). This conceptualisation offers an overall theoretical framework for vertical power that is necessary for sound competition law enforcement, and which has been lacking so far. The paper also explores specific metrics for vertical power, although this is still work in progress. Another important tool that competition authorities may employ in order to map the complex competitive interactions (horizontal and vertical) in the digital economy is the value chain approach. Although competition authorities have already used this tool in sector/industry inquiries, they have not in competition law adjudication. A value chain approach enables competition authorities to better assess the bargaining asymmetries across the various segments of the value chain that may result either from the lack of competition on the markets affected or from the central position of some actors in the specific network and their positioning in the value chain. This tool may complete the market definition tool. The effectiveness of competition law in the digital age may be curtailed by the cross-side network effects linked to positive feedback loops, increasing returns to scope and scale, the intense learning effects linked to AI, and the propensity of digital markets to tip. Hence, competition law on its own may not be sufficient to address the market failures in the digital economy. One therefore needs to take a toolkit approach that would combine different fields of law and regulation, competition law playing a primordial role in this new regulatory compass. This toolkit approach may rely on different combinations in each jurisdiction, on the basis of the institutional capabilities and the relative efficiency of the various regulatory alternatives, any choice being between imperfect, if perceived in isolation, institutional alternatives.	2019	Ioannis Lianos	Competition Law for the Digital Era: A Complex Systems’ Perspective
AI exclusion and exploitation Principles for AI Regulation and Competition Law	Algorithmic Lending, Competition, and Strategic Information Disclosure	Machine learning algorithms are increasingly used to evaluate borrower creditworthiness in financial lending, yet many lenders do not provide pre-approval tools that could significantly benefit consumers. These tools are essential for reducing consumer uncertainty and improving financial decision-making. This paper examines why symmetric lenders, with equal non-price features and algorithmic accuracy, might asymmetrically reveal pre-approval outcomes. Using a multi-stage game theory model, we analyze the strategic decisions of duopoly lenders in offering pre-approval tools for unsecured financial products. Our findings reveal that high algorithm accuracy can sustain an asymmetric revelation equilibrium, with one lender disclosing pre-approval outcomes while the other does not. Conversely, low algorithm accuracy prompts both lenders to reveal pre-approval outcomes. These findings diverge from traditional literature, which typically associates asymmetric revelation with differentiated products. Additionally, our results show that mandatory revelation policies could reduce lenders' incentives to improve algorithmic accuracy, potentially harming social welfare. These insights inform managerial strategies on the use of algorithmic transparency in lending and underscore the need for careful consideration of regulatory policies to balance market efficiency and consumer protection.	2023	Qiaochu Wang, Yan Huang, Param Vir Singh	Algorithmic Lending, Competition, and Strategic Information Disclosure
AI exclusion and exploitation Principles for AI Regulation and Competition Law	Algorithmic Personalized Pricing	Price is often the single most important term in consumer transactions. As the personalization of e-commerce continues to intensify, the law and policy implications of algorithmic personalized pricing i.e., to set prices based on consumers’ personal data with the objective of getting as closely as possible to their maximum willingness to pay (APP), should be top of mind for regulators. This article looks at the legality of APP from a personal data protection law perspective, by first presenting the general legal framework applicable to this commercial practice under competition and consumer law. There is value in analysing the legality of APP through how these bodies of law interact with one and the other. This article questions the legality of APP under personal data protection law, by its inability to effectively meet the substantive requirements of valid consent and reasonable purpose. Findings of illegality of APP under personal data protection law may in turn further inform the lawfulness of APP under competition and consumer law.	2020	Pascale Chapdelaine	Algorithmic Personalized Pricing
AI exclusion and exploitation Principles for AI Regulation and Competition Law	Dynamic Pricing Algorithms, Consumer Harm, and Regulatory Response	Pricing algorithms are rapidly transforming markets, from ride-sharing, to air travel, to online retail. Regulators and scholars have watched this development with a wary eye. Their focus so far has been on the potential for pricing algorithms to facilitate explicit and tacit collusion. This Article argues that the policy challenges pricing algorithms pose are far broader than collusive conduct. It demonstrates that algorithmic pricing can lead to higher prices for consumers in competitive markets and even in the absence of collusion. This consumer harm can be initiated by a single firm employing a superior pricing algorithm. Higher prices arise from the automated nature of algorithms, impacting any market where firms price algorithmically. Pricing algorithms that are already in widespread use may allow sellers to extract a massive amount of wealth from consumers. Because this consumer harm arises even when firms do not collude, antitrust law cannot solve the problem. This Article looks to the history of pricing innovation in the early twentieth century to show how government can respond when new pricing technologies and strategies disrupt markets. It argues for pricing regulation as a feasible solution to the challenges non-collusive algorithmic pricing poses, and it proposes interventions targeted at when and how firms set prices.	2021	Alexander MacKay, Samuel Weinstein	Dynamic Pricing Algorithms, Consumer Harm, and Regulatory Response
AI exclusion and exploitation Principles for AI Regulation and Competition Law	Algorithms and Fairness: What Role for Competition Law in Targeting Price Discrimination Towards End Consumers?	While algorithms bring about benefits for consumers in the form of more efficient price setting, concerns have also been expressed about possible adverse effects including discrimination. The paper takes a competition law perspective in analysing a type of discrimination that is said to be facilitated by the use of algorithms, namely personalised pricing. This is a form of price discrimination between consumers whereby a firm charges each consumer a different price depending on willingness to pay. As the advent of data analytics and algorithm-based services has made it easier for firms to engage in price discrimination, a clarification of the latter’s legality under competition law is to be welcomed. As such, the paper discusses the extent to which competition enforcement can be considered desirable to target price discrimination towards end consumers. In this regard, attention is also paid to the interaction with other regimes such as data protection, consumer protection and antidiscrimination law.	2018	Inge Graef	Algorithms and Fairness: What Role for Competition Law in Targeting Price Discrimination Towards End Consumers?
AI exclusion and exploitation AI and market power	Competition Law for the Digital Era: A Complex Systems’ Perspective	As the global economy incurs a process of transformation by the ongoing ‘fourth industrial revolution’, competition law is traversing a ‘liminal’ moment, a period of transition during which the normal limits to thought, self-understanding and behaviour are relaxed, opening the way to novelty and imagination, construction and destruction. There is need for the discussion over the role of competition law in the digital era to be integrated to the broader debate over the new processes of value generation and capture in the era of digital capitalism and the complex economy to which it has given rise to. This complex digital economy is formed by a spider web of economic links, but also their underpinning societal relations, between different agents. However, competition law still lives in the simple world of neo-classical price theory (NPT) economics, which may not provide adequate tools in order to fully comprehend the various dimensions of the competition game. The emphasis put recently by competition authorities on multi-sided markets in order to analyse restrictions of competition in the data economy illustrates the agents’ changing roles and the complexity of their interactions, as the same agents can be at the same time consumers and producers while their personal data raw material for the value generation process. It becomes therefore essential to uncover the new value capture and value generation processes in operation in the digital economy, and draw lessons for the optimal design and enforcement of competition law, rather than take the established competition law framework as a given and try to stretch within it a quite complex reality that may not fit this Procrustean iron bed. These approaches should engage with the complex economics of digital capitalism, and in particular the role of futurity and financialisation, personalisation and cybernetics. These new developments, first, call for a re-conceptualisation of the goals of competition law in the digital era, as competition law moves from the calm and predictable waters of ‘consumer welfare’, narrowly defined, to integrate considerations of income/wealth distribution, privacy and complex equality. Second, it also requires a revision of the current understanding of the nature of the competitive game, which only focuses on horizontal rivalry in product and eventually technology markets. This is of course an important dimension of competition, but hardly the most significant one in the current process of value generation and capture in the digital economy. Firms do not only compete on the product market dimension, but in the today’s financialised economy, probably the most important locus of competition is capital markets. The process of financialisation has important implications for the development of digital capitalism, an issue that the paper explores in detail for the first time in competition law and economics scholarship. Financial markets evaluate companies in view of expected returns in the not so near future, often linked to the emergence of bottlenecks or the perception that a firm holds important assets and resources (e.g. data, algorithms, specialised labour). The role of financial markets’ evaluation in driving business strategies in the era of digital and financialised capitalism is linked to the ‘subtle shift of mindset’ in digital capitalism ‘from profit (and isolating mechanisms) to wealth creation (and the potential for asset appreciation)’ as value is created by investing in assets that will appreciate. Third, this calls for a consideration, not only of horizontal competition, but also of vertical competition, the competition for a higher percentage of the surplus value brought by innovation, and competition from complementary technologies that may challenge the lead position in the value chain of the incumbents (vertical innovation competition). Fairness considerations, among other reasons, may also lead competition authorities to not only focus on inter-platform/ecosystem competition but to also promote intra-platform/ecosystem competition, as this may be a significant element of the competitive game. To implement this broader focus of competition law, we need to develop adequate conceptual tools and methodologies. A recurrent problem is the narrow definition of market power in competition law, whose presence often triggers the competition law assessment, and which is also intrinsically linked to the step of market definition. This currently ignores possible restrictions of vertical competition, personalisation and the predictive role of digital platforms, which may become source of harm for consumers, the competitive process, or the public at large. It is important to engage with concepts of vertical power and the paper develops a typology of vertical power, combining in an overall conceptual framework the various concepts of non-structural power that have been used so far in competition law literature and some new ones (positional and architectural power). This conceptualisation offers an overall theoretical framework for vertical power that is necessary for sound competition law enforcement, and which has been lacking so far. The paper also explores specific metrics for vertical power, although this is still work in progress. Another important tool that competition authorities may employ in order to map the complex competitive interactions (horizontal and vertical) in the digital economy is the value chain approach. Although competition authorities have already used this tool in sector/industry inquiries, they have not in competition law adjudication. A value chain approach enables competition authorities to better assess the bargaining asymmetries across the various segments of the value chain that may result either from the lack of competition on the markets affected or from the central position of some actors in the specific network and their positioning in the value chain. This tool may complete the market definition tool. The effectiveness of competition law in the digital age may be curtailed by the cross-side network effects linked to positive feedback loops, increasing returns to scope and scale, the intense learning effects linked to AI, and the propensity of digital markets to tip. Hence, competition law on its own may not be sufficient to address the market failures in the digital economy. One therefore needs to take a toolkit approach that would combine different fields of law and regulation, competition law playing a primordial role in this new regulatory compass. This toolkit approach may rely on different combinations in each jurisdiction, on the basis of the institutional capabilities and the relative efficiency of the various regulatory alternatives, any choice being between imperfect, if perceived in isolation, institutional alternatives	2019	Ioannis Lianos	Competition Law for the Digital Era: A Complex Systems’ Perspective
AI exclusion and exploitation AI facilitated manipulations AI and market power	Big Data and Personalised Price Discrimination in EU Competition Law	The networked digital revolution is ushering in a new data-driven age, powered by the engine of Big Data. We generate a massive volume of digital data in our everyday lives via our on-line interactions, which can now be tracked on a continuous and highly granular basis. Being able to track this data has radically disrupted the retail sector through, amongst other things, digital personalisation. However, this is no longer limited to shopping recommendations and advertising delivered to our smartphones, laptops and other mobile devices, but may extend to the prices at which goods and services are offered to customers in on-line environments, making it possible for two individuals to be offered exactly the same product, at precisely the same time, but at different prices, based on an algorithmic assessment of each shopper’s predicted willingness to pay. This is done by mining consumers’ digital footprints, using machine learning algorithms to enable digital retailers to predict the price that individual consumers (‘final end users’) are willing to pay for particular items, and thus offer them different prices. This phenomenon, which we dub ‘algorithmic consumer price discrimination’ (ACPD) forms the focus of this paper. The practice of price discrimination, which we define as “… charging different customers or different classes of customers different prices for goods or services whose costs are the same or, conversely, charging a single price to customers for whom supply costs differ…” is hardly a new phenomenon. Familiar forms include loyalty discounts, volume or multi-buy discounts, and the offering of status based discounts for students, old-age pensioners and the unemployed. However, the technological capacities of Big Data substantially enhance the ability of digital retailers to engage in much more precise, targeted and dynamic forms of price discrimination that were not previously possible. There are many areas of law that might mount a response to rising public anxieties associated with these practices. Our paper examines ACPD from the perspective of competition law through which we seek to evaluate ACPD by reference to two contrasting normative values: economic efficiency, on the one hand, and fairness or equity on the other. Competition law provides a unique lens for interrogating the social implications of ACPD due to its distinctive character as a form of ‘economic law’ that is intended to protect and strengthen the process of rivalry in the marketplace. Although ‘traditional’ forms of price discrimination have long been the subject of economic analysis to evaluate whether they are economically efficient, algorithmic price discrimination has hitherto attracted relatively little critical analysis. As we demonstrate in Section 2, the incentives for firms to engage in ACPD often exist. We find that consumers are in the aggregate often better off, economically, when sellers can price discriminate in this way, thereby enhancing consumer surplus. However, this is not always the case. Furthermore, whether EU competition law is solely and exclusively concerned with economic efficiency, or whether it provides scope for non-efficiency based considerations in the application of its provisions, is a matter of debate. Accordingly, in Section 3 we evaluate ACPD by reference to its fairness or justice (which we also call equity) understood in three distinct (and sometimes overlapping) ways: (a) the perceived fairness of pricing practices; (b) unfair dealing between online retailers and consumers (corrective justice); and (c) fairness as a requirement of distributive (or collective) justice. For each of these understandings of fairness, we identify points of convergence and conflict with economic evaluations of the effects of ACPD on aggregate consumer welfare. No Article 102 cases have directly considered the legality of ACPD. Section 4 therefore interrogates existing Article 102 case law to ascertain whether ACPD would likely breach this provision. Because the current legal position is unclear, Section 5 draws together the efficiency and fairness evaluations by considering whether ACPD should be regarded as unlawful under EU competition law. We argue that where ACPD increases both consumer surplus and fairness, it should not breach Article 102. Conversely, where ACPD undermines both consumer welfare and fairness, then such practices should be unlawful under Article 102. However, because economic and fairness evaluations of ACPD may conflict in specific cases, Section 5 also considers whether, in the light of the underlying justifications for EU competition law and the EU’s foundational principles, ACPD should be considered a violation of Article 102 where it undermines justice or equity, even though it may enhance consumer surplus, and vice versa. We deal with the clashes between these goals in two ways: first, we offer a partial reconciliation between these goals, by supplementing conventional economic analysis with insights from behavioural economics, thus enabling some fairness considerations that affect consumer welfare to be taken into account. Secondly, we suggest that fairness should have a secondary role when Article 102 is applied to ACPD, in the form of a ‘defence’ to an allegation of abuse of market power. On our suggested account, ACPD which reduces consumer surplus may nonetheless avoid falling foul of Article 102 if it can be justified on grounds of fairness. Section 6 concludes, suggesting that EU competition law may have a valuable but limited role to play in redressing some of the adverse impacts of ACPD, primarily by focusing on the consumer welfare effects of ACPD, and in which considerations of fairness and justice play a relevant, but nonetheless subsidiary, role. Competition law cannot, and should not, seek to solve all the social problems associated with market behaviour, including data-driven forms of personalised pricing	2017	Christopher Townley, Eric Morrison, Karen Yeung	Big Data and Personalised Price Discrimination in EU Competition Law
AI exclusion and exploitation AI and market power	Data-Driven Economy and Artificial Intelligence: Emerging Competition Law Issues	Artificial Intelligence (AI), although it is not an entirely new phenomenon, has been developing more intensively relatively recently. AI is based on artificial neural networks (ANNs), which are modelled upon the architecture of human brains. Beside several ethical issues, such as whether “artificial brains” can and should be created in the first place, a number of legal questions emerge. One of the most urgent issues to be analyzed as regards AI is whether there is a need for a regulatory intervention. It is probably fair to say that AI is currently in a shape of a cocoon. In an academic legal research AI is still pretty much “terra incognita”. However, although there is no clear and visible AI market failure (yet), it may be worthwhile analyzing what the status quo of the markets in the digital economy is with a perspective view of the AI markets of the future. After all, AI is featured by the systems that can autonomously learn and improve (machine learning). Such self-learning capability of the machines is based on the technique called “deep learning”. The latter is fed by data. In this regard, it may well be that the purpose of current data collection and processing is related to conquering future markets for AI, so that a “snapshot” analysis of the issues related to current data-driven markets would show only one side of the coin of the market dynamics. The latter insight has to be borne in mind by both regulators and competition authorities. It is in this context that it has to be analyzed whether or what type of protection is needed for (non-personal) data and what is the optimal scope of protection of “deep learning” algorithms. After all, a current “open source” strategy of the biggest market players may be attractive from a short-run perspective, but may possibly raise competition law concerns in a long run. If current processes in the markets feature competition for developing future systems of AI, the question of (setting) standards and interoperability may turn out to be mostly important for future competition.	2017	Gintare Surblyte	Data-Driven Economy and Artificial Intelligence: Emerging Competition Law Issues
AI exclusion and exploitation AI and market power	Artificial Intelligence, Algorithmic Recommendations and Competition	We present a methodology for analyzing the impact of algorithmic recommendations on product market competition, addressing concerns that have been raised in both academic and policy circles regarding their potential anti-competitive effects. Our analysis demonstrates that recommender systems (RSs) lead to higher market concentration and prices compared to a scenario where algorithmic recommendations are unavailable and consumers rely solely on individual search. However, RSs also improve the match between products and consumers and reduce the need for expensive search processes. By accounting for both the positive and negative effects, we find that RSs are likely to increase consumer surplus for reasonable parameter values. However, increasing the amount of data available to the algorithms may lead to a reduction in consumer surplus. We also examine the potential for manipulation of recommendations and its impact on competition, finding that such manipulation is more likely to represent an exclusionary abuse than an exploitative one.	2025	Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò, Sergio Pastorello	Artificial Intelligence, Algorithmic Recommendations and Competition
AI exclusion and exploitation AI facilitated manipulation	Algorithmic Challenges to Autonomous Choice	Human choice is a foundational part of our social, economic and political institutions. This focus is about to be significantly challenged. Technological advances in data collection, data science, artificial intelligence, and communications systems are ushering in a new era in which digital agents, operated through algorithms, replace human choice with regard to many transactions and actions. While algorithms will be given assignments, they will autonomously determine how to carry them out. This game-changing technological development goes to the heart of autonomous human choice. It is therefore time to determine whether and, if so, under which conditions, are we willing to give up our autonomous choice. To do so, this article explores the rationales that stand at the basis of human choice, and how they are affected by autonomous algorithmic assistants; it conscientiously contends with the “choice paradox” which arises from the fact that the decision to turn over one’s choices to an algorithm is, itself, an act of choice. As shown, while some rationales are not harmed – and might even be strengthened – by the use of autonomous algorithmic assistants, others require us to think hard about the meaning and the role that choice plays in our lives. The article then examines whether the existing legal framework is sufficiently potent to deal with this brave new world, or whether we need new regulatory tools. In particular, it identifies and analyzes three main areas which are based on choice: consent, intent and laws protecting negative freedom.	2017	Michal Gal	Algorithmic Challenges to Autonomous Choice
AI exclusion and exploitation AI and market power	Big Data and Emerging Competition Concerns	This paper identifies access to Big Data as one of the important factors for the success and growth of online platforms. Through Big Data, businesses can track market trends and use target advertising services in ways that were previously impossible. The data can be leveraged to increase market power through a number of artificial intelligence-based advances, thereby increases barriers to entry in the relevant market. Dominant online platforms can use Big Data to enter into certain anti-competitive acts such as price discrimination as well as refuse access to data which can enhance barriers to entry in the relevant market. Hence, this paper seeks to examine the above-mentioned competition concerns and their possible remedies under competition law.	2021	Aaqib Javeed	Big Data and Emerging Competition Concerns
AI exclusion and exploitation AI and market power	The Faux Pas in Modern Competition Law – Walled Gardens, Data Sharing and Algorithmic Decision Making	The milieu of the 21st century has triggered a wave of unprecedented changes across traditional market structures, igniting disruption and necessitating evolution in firms big and small. A brief survey of the current global climate reveals the digital economy largely requiring some form of intervention – lest market abuse arise to the detriment of the modern consumer. In the United States, the Gordian Knot of walled gardens in the social media industry has triggered antitrust attention; where ‘Google’ and ‘Facebook’, juggernauts of the social media industry, have largely created a confined duopoly system. Indeed, the ability for said companies to access much sought-after consumer data, led to regulation being necessary to prevent market abuse. Winging this issue to the United Kingdom and the European Union, technological developments have led to a necessary change in regulations – to facilitate innovation, while at the same time to ensure adequate consumer protection. This paper will adopt a two-pronged approach – in the first part, an economics-focused view will be adopted to examine the present digital economy; and in the second - the current regime in the UK will be analysed from a legal perspective, focusing on how Art 101 TFEU and Chapter I of the UK Competition Act affects firms from a top-down level. The final scope of this argument contrasts the Bundeskartellamt’s investigation into Facebook with AGCM’s investigation in Italy, fully fleshing out the regulatory dilemmas encountered by competition authorities of the region. In the final analysis, this paper argues that more governmental intervention is required in three sub-areas; namely in (i) data sharing, (ii) self-learning algorithms and finally, (iii) marketplace(s) with walled gardens	2021	Kan Jie Marcus Ho	The Faux Pas in Modern Competition Law – Walled Gardens, Data Sharing and Algorithmic Decision Making

AI facilitated collusion

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
AI facilitated collusion AI andMarket Power	Algorithmic Collusion in the Skies: The Role of AI in Shaping Airline Competition	Since the mid-2010s, U.S. airlines have implemented and expanded the use of artificial intelligence (AI) in a variety of capacities. Given the industry's oligopolistic market structure dominated by a few major carriers, AI-driven pricing algorithms may raise antitrust concerns due to their potential to facilitate coordinated pricing and tacit collusion. This paper reviews recent developments in the economic literature on these issues and examines the mechanisms through which AI-driven pricing algorithms may enable coordinated pricing among airlines. It also surveys existing experimental and empirical evidence on the competitive effects of AI adoption and explores the regulatory challenges posed by algorithmic coordination in the airline industry. The paper concludes with policy recommendations to mitigate potential anticompetitive risks.	2025	Qi Ge, Myongjin Kim, Nicholas G Rupp	Algorithmic Collusion in the Skies: The Role of AI in Shaping Airline Competition
AI facilitated collusion AI mergers and cooperation	Limiting Algorithmic Coordination	Recent studies have proven that pricing algorithms can autonomously learn to coordinate prices, and set them at supra-competitive levels. The growing use of such algorithms mandates the creation of solutions that limit the negative welfare effects of algorithmic coordination. Unfortunately, to date, no good means exist to limit such conduct. While this challenge has recently prompted scholars from around the world propose different solutions, many suggestions are inefficient or impractical, and some might even strengthen coordination. This challenge requires thinking outside the box. Accordingly, this article suggests four (partial) solutions. The first is market-based, and entails using consumer algorithms to counteract at least some of the negative effects of algorithmic coordination. By creating buyer power, such algorithms can also enable offline transactions, eliminating the online transparency that strengthens coordination. The second suggestion is to change merger review so as to limit mergers that are likely to increase algorithmic coordination. The next two are more radical, yet can capture more cases of such conduct. The third involves the introduction of a disruptive algorithm, which would disrupt algorithmic coordination by creating noise on the supply side. The final suggestion entails freezing the price of one competitor, in line with prior suggestions to address predatory pricing suggested by Edlin and others. The advantages and risks of each solution are discussed. As antitrust agencies around the world are just starting to experiment with different ways to limit algorithmic coordination, there is no better time to explore how best to achieve this important task.	2022	Michal Gal	Limiting Algorithmic Coordination
AI facilitated collusion	Algorithmic collusion under competitive design	I study a model in which two players (designers) simultaneously choose an exploration policy for their Q-learning algorithms. Their algorithms then repeatedly play a stage game chosen from a class including prisoner’s dilemma, first and second price auctions and Bertrand competition. Designers collect the limiting payoffs obtained by their algorithms. In equilibrium, both players always receive payoffs that are strictly higher than the payoffs received in the unique strict Nash equilibrium of the stage game. A restriction of this model to the prisoner’s dilemma played by ε-greedy Q-learning algorithms is also studied. I perform extensive numerical simulations that allow to gain insight on (i) the mechanism causing algorithmic collusion and (ii) the strategic role of exploration levels in the game. These findings have implications for the regulation of algorithmic collusion.	2025	Ivan Conjeaud	Algorithmic collusion under competitive design
AI facilitated collusion AI exclusion and exploitation	Algorithmic Collusion and Price Discrimination: The Over-Usage of Data	As firms' pricing strategies increasingly rely on algorithms, two concerns have received much attention: algorithmic tacit collusion and price discrimination. This paper investigates the interaction between these two issues through simulations. In each period, a new buyer arrives with independently and identically distributed willingness to pay (WTP), and each firm, observing private signals about WTP, adopts Q-learning algorithms to set prices. We document two novel mechanisms that lead to collusive outcomes. Under asymmetric information, the algorithm with information advantage adopts a Bait-and-Restrained-Exploit strategy, surrendering profits on some signals by setting higher prices, while exploiting limited profits on the remaining signals by setting much lower prices. Under a symmetric information structure, competition on some signals facilitates convergence to supra-competitive prices on the remaining signals. Algorithms tend to collude more on signals with higher expected WTP. Both uncertainty and the lack of correlated signals exacerbate the degree of collusion, thereby reducing both consumer surplus and social welfare. A key implication is that the over-usage of data, both payoff-relevant and non-relevant, by AIs in competitive contexts will reduce the degree of collusion and consequently lead to a decline in industry profits.	2024	Zhang Xu, Mingsheng Zhang, Wei Zhao	Algorithmic Collusion and Price Discrimination: The Over-Usage of Data
AI facilitated collusion	Human-Algorithm Interaction: Algorithmic Pricing in E Laboratory Markets	This paper investigates pricing in laboratory markets when human players interact with an algorithm. We compare the degree of competition when exclusively humans interact to the case of one firm delegating its decisions to an algorithm, an n -player generalization of tit-for-tat. We further vary whether participants know about the presence of the algorithm. When one of three firms in a market is an algorithm, we observe significantly higher prices compared to human-only markets. Firms employing an algorithm earn significantly less profit than their rivals. (Un)certainty about the actual presence of an algorithm does not significantly affect collusion, although humans seem to perceive algorithms as more disruptive.	2022	Hans-Theo Normann, Martin Sternberg	Human-Algorithm Interaction: Algorithmic Pricing in Hybrid Laboratory Markets
AI facilitated collusion	Pricing via Artificial Intelligence: The Impact of Neural Network Architecture on Algorithmic Collusion	Classic artificial intelligence (Q-learning) algorithms have been capable of consistently learning supra-competitive pricing strategies in infinitely repeated Nash-Bertrand pricing games without human communication. Such algorithms have been able to converge due to the temporal correlation of consecutive states and actions in the learning process, which restores stationarity in an otherwise highly non-stationary setting. It is difficult for more realistic AI algorithms to converge, as the necessary training processes breaks the aforementioned temporal correlation, rendering the algorithms ineffective in learning reward-punishment strategies that result in collusive market outcomes. We adapt several widely used neural network architectures to the framework of model-free reinforcement learning and experimentally explore how the structure of AI algorithms affects market outcomes in a workhorse oligopolistic model of repeated price competition. While it is possible to train advance AI algorithms to always best respond in environments where the rival exercises a fixed strategy, it is unlikely that such algorithms can learn to coordinate in setting supra-competitive prices due to the non-stationarity of multi-agent learning processes, suggesting that algorithmic collusion may not be an immediate concern for antitrust authorities.	2023	Weipeng Zhang	Pricing via Artificial Intelligence: The Impact of Neural Network Architecture on Algorithmic Collusion
AI facilitated coordination and collusion	Collusive Outcomes Without Collusion: Algorithmic Pricing in a Duopoly Model	We develop a model of algorithmic pricing which shuts down every channel for explicit or implicit collusion, and yet still generates collusive outcomes. We analyze the dynamics of a duopoly market where both firms use pricing algorithms consisting of a parameterized family of model specifications. The firms update both the parameters and the weights on models to adapt endogenously to market outcomes. We show that the market experiences recurrent episodes where both firms set prices at collusive levels. We analytically characterize the dynamics of the model, using large deviation theory to explain the recurrent episodes of collusive outcomes. Our results show that collusive outcomes may be a recurrent feature of algorithmic environments with complementarities and endogenous adaptation, providing a challenge for competition policy.	2024	In-Koo Cho, Noah Williams	Collusive Outcomes Without Collusion: Algorithmic Pricing in a Duopoly Model
AI facilitated collusion AI and pricing	Autonomous pricing using policy gradient reinforcement learning	Artificial intelligence algorithms are increasingly used by firms to set prices. Previous research on pricing algorithms shows that they can exhibit collusive behavior, but it has so far remained an open question whether they can do so in a reasonably short time. I develop a deep reinforcement learning model able to price goods in a repeated oligopolistic competition game with continuous prices that, under reasonable assumptions on the length of a time step, converges to a collusive outcome in an amount of time that matches empirical observations. The model I propose reliably shows cooperative behavior supported by reward-punishment schemes that discourage deviations from the point of convergence.	2023	Kevin Michael Frick	Autonomous pricing using policy gradient reinforcement learning
AI facilitated collusion AI and Market Power Exclusion and exploitation AI and pricing	Fixing Algorithmic Pricing? Competition Concerns and Solutions	Pricing algorithms can leverage a litany of inputs, including costs, competitor pricing, market supply, inventory, production levels, capacity constraints, consumer demand, efficiencies, and business objectives. Like many technological transitions, the rapid advancement and widespread adoption of pricing algorithms can raise important questions and implicate significant tradeoffs, Part I of this article explains the legal framework for evaluating claims of collusion under Section 1 of the Sherman Antitrust Act. Part II analyzes the three core competitive concerns around algorithmic pricing, including recent antitrust litigation reflecting these issues. Part III discusses the various legislative proposals to address these concerns. Part IV offers my conclusion that each of the three flavors of competitive concerns discussed in Part II requires a solution that is narrowly tailored. Stated differently, a one-size-fits-all approach is not an optimal solution.	2025	Henry Hauser	Fixing Algorithmic Pricing? Competition Concerns and Solutions
AI facilitated collusion	Algoritmische besluitvorming en het kartelverbod (Algorithmic Decion-Making and Antitrust Law)	English Abstract: Algoritms increasingly take over decision-making from human actors. Using big data (large volume of usually unstructured and realtime data), algorithms can react faster and better than human actors. The use of algoritmic decision-making may have competition law effects on markets and can,for instance in cases of automated or personalised pricing, lead to (new types of) cartels. This paper focuses on the technological developments and potential consequences in antitrust law.	2018	Anna Gerbrandy, Bart Custers	Algoritmische besluitvorming en het kartelverbod (Algorithmic Decion-Making and Antitrust Law)
AI facilitated collusion	Anti-Competitive Nature of Pricing Algorithms	The Pricing strategy is one of the major determinants of the success or failure of any business endeavor. Such determinants are made through algorithms virtually now-a-days, which causes vertical agreement, which is not per se anti-competitive in nature. The advancement of 'pricing algorithms' presents one such issue. This becomes anti-competitive only when the pricing structure or strategy violates Section 3 of the Competition Act, 2002. Price fixing per se cannot be anti-competitive in nature. There could be price-fixing either traditionally or dynamically. The competition act aims in giving complete protection to the consumers and other competitors and prevent them from being exploited due to anti-competitive practices or any other unfair or unhealthy trade practices. The dynamic methods of price-fixing are evolving with greater importance nowadays.	2021	Hariharan Venkateshwaran	Anti-Competitive Nature of Pricing Algorithms
AI facilitated collusion	Algorithmic Pricing: Implications for Marketing Strategy and Regulation	Over the past decade, a growing number of firms have delegated pricing decisions to algorithms in consumer and business markets such as travel, entertainment, and retail, as well as in platform markets such as ride-sharing. We define algorithmic pricing as “the use of programs to automate the setting of prices.” Firms adopt algorithmic pricing to optimize their prices in response to changing market conditions and to leverage the efficiency gains from automation. Advances in information technology and the increased availability of digital data have further facilitated the use of algorithm-driven pricing strategies. Yet adopting algorithmic pricing is not merely a technical upgrade--it is a strategic decision that must align with a company's existing and future marketing strategies. Moreover, algorithmic pricing can raise various regulatory concerns regarding potential threats to competition and the legality of price discrimination. This paper discusses the implementation of algorithmic pricing in the context of firms' marketing strategies and regulatory frameworks, while outlining an agenda for future research in this increasingly important area.	2025	Martin Spann, Marco Bertini, Oded Koenigsberg, Robert Zeithammer, Diego Aparicio, Yuxin Chen, Fabrizio Fantini, Ginger Zhe Jin, Vicki Morwitz, Peter T. L. Popkowski Leszczyc, Maria Ana Vitorino, Gizem Yalcin Williams, Hyesung Yoo	Algorithmic Pricing: Implications for Marketing Strategy and Regulation
AI facilitated collusion	How Much Is No Longer A Simple Question – Pricing Algorithms and Antitrust Laws	With the backdrop of increased use of algorithms in doing business, the objective of this paper is examining in-depth the use of one such category of algorithms – pricing algorithms. The scope of this paper is confined to two potential issues associated with the use of pricing algorithms – algorithmic collusion and personalized pricing. Finally, this paper offers solutions that may be used by antitrust agencies in the United States, i.e., the Department of Justice and the Federal Trade Commission, to deal with these issues within the confines of the existing antitrust jurisprudence. Section I of this paper describes what pricing algorithms are, and how they assist with dynamic pricing. The section demonstrates that it is irrefutable that the use of pricing algorithms has improved the competitive landscape by increasing transparency, reducing information asymmetry and improving the allocative efficiency of the market, bringing it closer to the ideal of perfect competition. At the same time, agencies are wary that the use of big data analytics and pricing algorithms could lead to the anti competitive outcomes which both contribute to and are aggravated by concentration of market power. Section II examines the role pricing algorithms play in facilitating in explicit and tacit collusion. The section also concludes that the existing framework of antitrust law would be sufficient while dealing with algorithmic forms of explicit collusion, where an algorithm is used to execute an anti competitive agreement but may not suffice for tackling algorithmic tacit collusion. Further, trying to uncover evidence of anti-competitive agreement and intent may prove challenging as machine learning is an ongoing process and the design of an algorithm can be complicated to decipher. Section III explores the use of algorithms in personalizing pricing. This section notes that while from an economic perspective, personalized pricing fosters static efficiency, it also has the tendency to be used to exploit unsuspecting consumers. Policymakers and antitrust agencies are also concerned about issues of fairness and consumer welfare regarding the use of pricing algorithms. However, given that the unilateral use of a pricing algorithm to set high prices is not an offense under US antitrust laws, agencies are skeptical of prosecuting the use of exploitative personalized pricing. Section IV puts forward some possible solutions. The paper reasons that at this time, rather than making legislative changes, it is essential for the FTC to invest in research into such methods in order to both uncover evidence of anti-competitive agreements and audit algorithms that appear to collude even in the absence of an agreement. Further, this paper argues that FTC may have the locus to investigate exploitative personalized pricing under Section 5 of the FTC Act . Though, the most practical short-term solution seems to be to increase transparency and to put power back into the hands of the consumer.	2018	Ankita Gulati	How Much Is No Longer A Simple Question – Pricing Algorithms and Antitrust Laws
AI facilitated collusion	Pricing Algorithms—What Role For Regulation?	The rapid spread of pricing algorithms in e-commerce markets has raised alarms about their potential for anticompetitive abuse. Enforcers and policy-makers have been concerned for some time about the possibility of widespread algorithmic price-fixing and dominant firms’ use of algorithms to damage rivals. These harms are in theory redressable under the antitrust laws. But evidence is mounting that pricing algorithms will raise prices to consumers in ways that do not violate the antitrust laws. Tacit algorithmic collusion and price increases due to competition among pricing algorithms will make many online goods and services more expensive. Consumers currently have no effective way to fight back against these higher prices. Market-based solutions, like consumer-friendly algorithms that steer buyers to the best prices, can help. But considering the scope and scale of the ongoing revolution in pricing technology, protecting consumers is likely to require a regulatory response. Regulations that limit when and how firms set prices could restrict algorithms’ ability to raise prices above the competitive level. While not costless, this approach might be necessary to prevent a significant transfer of wealth from consumers to sellers.	2024	Samuel Weinstein	Pricing Algorithms—What Role For Regulation?
AI facilitated collusion	Explainable AI and Pricing Algorithms: A Case For Accountability in Pricing	As artificial intelligence (AI) continues to shape various industries, concerns arise regarding its potential anti-competitive implications. This research paper delves into the intersection of AI and competition law, with a specific focus on the role of explainability in identifying and addressing anti-competitive AI practices under the purview of Indian competition law. The paper begins by analyzing the challenges posed by the lack of transparency and explainability in AI models deployed by companies in market competition by analyzing 3 possibilities. It highlights the risks associated with opaque AI algorithms regarding tacit collusion that hinder the detection of anti competitive behavior, limit accountability, and impede fair competition. It also analyzes the possibility of pro-competitive behavior using algorithms. To propose solutions, the paper advocates for enhanced transparency and algorithmic accountability, emphasizing the need for companies to disclose information about their AI algorithms and justify their decisions. It highlights the need for antitrust authorities to develop expertise in assessing AI algorithms and enforcing competition laws effectively by assessing EU’s Competition and market authority and India’s Competition commission and their experiences in dealing with digital antitrust issues. Overall, this research aims to contribute to the evolving field of AI and competition law by shedding light on the significance of explainability, proposing solutions to address anti-competitive AI practices, and considering the specific context of Competition law. By ensuring transparency, accountability, and fairness in AI systems, we can promote competition, protect consumer welfare, and foster innovation in the AI-driven marketplace.	2023	Brahm Sareen	Explainable AI and Pricing Algorithms: A Case For Accountability in Pricing
AI facilitated coordination and collusion	Algorithms and Conscious Parallelism: Why Current Antitrust Doctrine is Prepared for the Twenty-First Century Challenges Posed by Dynamic Pricing	As the 21st Century enters its third decade, antitrust laws that are currently in place in the United States must confront an explosion in technological innovation. For many, this explosion is welcome news and will likely lead to a prosperous future for consumers. For others, including Professors Ariel Ezrachi and Maurice Stucke, they believe this future should be perceived cautiously and new antitrust laws and regulations should be erected to replace the old. They contend in their book, VIRTUAL COMPETITION: THE PROMISE AND PERILS OF THE ALGORITHM DRIVEN ECONOMY, coupled with their recent law review article, Sustainable and Unchallenged Algorithmic Tacit Collusion, that in the (near) future, algorithms will be able to collude with one another and modern day antitrust doctrine is not suited to counteract this virulently offensive and illegal behavior. Specifically, they contend that conscious parallelism poses a significant danger to competition policies and new doctrine should be developed to counter it. This article argues that the United States’ conscious parallelism plus factors maintain their value to antitrust regulators and courts even in this highly technical, rapidly evolving environment facilitated by technology. Ezrachi & Stucke’s arguments to the contrary are actually belied by the history of the doctrine, the technology itself, and their own arguments and descriptions of the technology. Furthermore, while some European empirical research sides with Ezrachi & Stucke, the competition authorities and other scholars sufficiently rebut these arguments by pointing out these experiments fail to reflect real world environments. All in all, we should approach this dynamic pricing cautiously and require oversight that is already constructed by regulators over corporations. However, antitrust authorities and courts should not create new laws to combat problems that have already been solved.	2020	John Fortin	Algorithms and Conscious Parallelism: Why Current Antitrust Doctrine is Prepared for the Twenty-First Century Challenges Posed by Dynamic Pricing
AI facilitated coordination and collusion	Strategic Choice of Price-Setting Algorithms	Recent experimental simulations have shown that autonomous pricing algorithms are able to learn collusive behavior and thus charge supra-competitive prices without being explicitly programmed to do so. These simulations assume, however, that both firms employ the identical price-setting algorithm based on Q-Learning. Thus, the question arises whether the underlying assumption that both firms employ this type of algorithm can be supported as an equilibrium if firms can choose between different pricing rules. Our simulations show that when both firms use an algorithm based on Q-learning, the outcome is not an equilibrium if alternative pricing rules are available. In fact, simpler pricing rules as for example meeting competition clauses yield higher payoffs compared to Q-learning algorithms.	2023	Ulrich Schwalbe, Katrin Buchali, Jens Grüb, Muijs Matthias	Strategic Choice of Price-Setting Algorithms
AI facilitated coordination and collusion	Controlling Algorithmic Collusion: Short Review of the Literature, Undecidability, and Alternative Approaches	Algorithms have played an increasingly important role in economic activity, as they becoming faster and smarter. Together with the increasing use of ever larger data sets, they may lead to significant changes in the way markets work. These developments have been raising concerns not only over the rights to privacy and consumers’ autonomy, but also on competition. Infringements of antitrust laws involving the use of algorithms have occurred in the past. However, current concerns are of a different nature as they relate to the role algorithms can play as facilitators of collusive behavior in repeated games, and the role increasingly sophisticated algorithms can play as autonomous implementers of pricing strategies, learning to collude without any explicit instructions provided by human agents. In particular, it is recognized that the use of ‘learning algorithms’ can facilitate tacit collusion and lead to an increased blurring of borders between tacit and explicit collusion. Several authors who have addressed the possibilities for achieving tacit collusion equilibrium outcomes by algorithms interacting autonomously, have also considered some form of ex-ante assessment and regulation over the type of algorithms used by firms. By using well-known results in the theory of computation, I show that such option faces serious challenges to its effectiveness due to undecidability results. Ex-post assessment may be constrained as well. Notwithstanding several challenges face by current software testing methodologies, competition law enforcement and policy have much to gain from an interdisciplinary collaboration with computer science and mathematics.	2018	Joao E Gata	Controlling Algorithmic Collusion: Short Review of the Literature, Undecidability, and Alternative Approaches
AI facilitated coordination and collusion	Should the EU Reshape its Competition Legal System to Regulate Algorithmic Cartels?	AI has undoubtedly brought beneficial innovations, but from a regulatory perspective, it is alleged that it challenges the antitrust system. In this regard, this Paper shall critically assess the issues which robots on AI pose on Article 101 and whether this framework is up to the task in regulating them. That is because, some academics and practitioners concur that in the event if algorithms start fixing prices by themselves without the involvement on human intervention, the virtual ‘meeting of minds’ is not likely to be caught by current case law or legislation due to the perfect forms of collusion they can reach. Also, in AI cases, there is no legal basis to attribute liability, which leaves the question open as to whom liability is attached should things go wrong – is it the programmer of the algorithms, its user or both? The discussion – where appropriate – will be supplemented with academic research and surveys. The starting point presents the problems related to the notion of Agreement under 101 and the oligopoly defense; the evidence of intent and horizontal agreements; the liability aspect; and the three approaches towards tacit algorithmic collusion. Finally, the discussion will shift on the possible applicable reforms and regulatory approaches, followed be a conclusion.	2018	Francisc Ioanid Toma	Should the EU Reshape its Competition Legal System to Regulate Algorithmic Cartels?
AI facilitated coordination and collusion	Collusion via algorithms: Comparative analysis and the way forward	The technological advancement at an unprecedented pace has resulted in legislation and regulation failing to keep up with new challenges faced by the consumers. The legislators and regulators are consequently struggling to amend provisions. In the traditional sense, collusion is ‘amongst’ humans or actions caused by humans. New technological advances have resulted in such collusions being taken by or in active connivance of algorithms. The major issues across jurisdictions are: 1) What constitutes collusion in the digital age?; and 2) What is the method to collect such evidence that they can rely on to enforce Competition law without disclosing the proprietary commercially sensitive information and scuttling innovation? The aim and objective of this paper is to analyze the functioning of algorithms under the lens of competition law, to draw a comparative analysis of the draft regulations from the jurisdictions of the European Union, the United Kingdom, and the United States, and provide suggestive regulatory reforms in India by analyzing the regulations being implemented with regard to algorithmic collusions in the abovementioned jurisdictions. The paper shall also attempt to put forward other measures (which are not a part of any regulation yet) that can be implemented to counter collusion by algorithms.	2021	Mohsin Rahim	Collusion via algorithms: Comparative analysis and the way forward
AI facilitated coordination and collusion	Algorithmic Pricing – A Black Box for Antitrust Analysis	The conversation around and study of the use of algorithms in pricing and other competitively sensitive decisions remains vibrant and is increasingly well-informed. Early theoretical work paved the way for government studies and more recently – and most interestingly – experimental and real-world empirical studies. At the same time, technology continues to advance, and with it the varieties and sophistication of software deployed. The law does not seem to have kept pace. Examples of enforcement to date are against pure cartel agreements that happen to have pricing algorithms as a tool for implementation. The most likely harms from deployment of pricing algorithms, increased capacity for optimal tacitly collusive outcomes, is unlikely to violate the law in any developed antitrust system. More speculative harms, including actual algorithmic collusion, seem to be equally outside of the realm of antitrust. And all of these considerations arise against a backdrop of efficiency considerations that while apparent seem to be under-theorized and under-studied. We outline findings on algorithmic pricing in theoretical and empirical research, how they interact with existing legal rules, and suggest promising areas for future study and policy development.	2022	Max Huffman, Maria José Schmidt-Kessen	Algorithmic Pricing – A Black Box for Antitrust Analysis
AI facilitated coordination and collusion	Algorithm-Based Pricing in Online Retailing As Concerted Practice Covering the ‘Predictable Agent’ With Article 101 (1) TFEU	Algorithm-supported pricing in online retailing markets is increasingly changing the conditions of competition. The enormous speed of price competition in the highly transparent markets of the digital economy, where companies can react significantly faster than consumers, promotes interdependent behavioural strategies between competitors with a tendency towards supracompetitive prices, even in markets that are not oligopolistically structured. Therefore, criteria are worked out in this article under which companies leave the area of market-induced parallel behaviour and instead the use of automated price algorithms can be seen as concerted behaviour within the meaning of Article 101 (1) TFEU.	2019	Wolf Maik	Algorithm-Based Pricing in Online Retailing As Concerted Practice Covering the ‘Predictable Agent’ With Article 101 (1) TFEU
AI facilitated coordination and collusion AI mergers and cooperation	Gig Platforms as Hub-and-Spoke Arrangements and Algorithmic Pricing: A Comparative EU-US Antitrust Analysis	Gig platforms are a modern economy enterprise structure characterized by a firm matching service providers with consumers – prominent examples include ride-sharing platforms, like Uber; delivery platforms, like Wolt; and lodging rental platforms, like Airbnb. As all online platforms, gig platforms are data-driven business models that employ and develop algorithms and AI tools that learn from user behavior and adapt to make interactions increasingly efficient. In contrast to other online platforms, such as advertising exchanges or online market places for goods, gig platforms enable users to sell their labor or services to other users via the platform. Scholarship has shown enterprises that contracts with their service providers, who are then by necessity operating as independent enterprises, are best analyzed as agreements implicating Art. 101 TFEU and Section 1 of the Sherman Act. Currently, the dominant legal treatment of service providers on platforms including Uber (ride-sharing) and Wolt (food delivery) is as contractors rather than employees. We employ here the lens of a hub-and-spoke arrangement, with the platform as the hub and the service providers as the spokes, and the algorithmically-established price terms representing a collection of parallel vertical agreements. We then engage in a comparative study of the legal implications under antitrust law in the US and the EU of hub-and-spoke arrangements. The chapter proceeds to outline the hub-and-spoke structure of the service provider-platform agreements in a gig economy enterprise, including the universal agreement to abide by prices set by algorithm in contracting for services. It covers various design options for pricing algorithms that can be used by platforms to coordinate the transaction between its users. Next, the chapter considers the EU caselaw on hub-and-spoke arrangements, analyzing authorities from across the EU, and identifies the probable treatment of the gig economy agreements in the light of these authorities. The chapter then conducts a similar analysis of leading recent authorities in the US and likewise concludes the most probable treatment under US law. In the conclusion, the chapter compares and explains the likely legal treatment of an algorithmically defined hub-and-spoke agreement and suggests areas for change.	2021	Max Huffman, Maria José Schmidt-Kessen	Gig Platforms as Hub-and-Spoke Arrangements and Algorithmic Pricing: A Comparative EU-US Antitrust Analysis
AI facilitated coordination and collusion AI mergers and cooperation	Collusion by Mistake: Does Algorithmic Sophistication Drive Supra-Competitive Profits?	A burgeoning literature shows that self-learning algorithms may, under some conditions, reach seemingly-collusive outcomes: after repeated interaction, competing algorithms earn supra-competitive profits, at the expense of efficiency and consumer welfare. This paper offers evidence that such behavior can stem from insufficient exploration during the learning process and that algorithmic sophistication might increase competition. In particular, we show that allowing for more thorough exploration does lead otherwise seemingly-collusive Q-learning algorithms to play more competitively. We first provide a theoretical illustration of this phenomenon by analyzing the competition between two stylized Q-learning algorithms in a Prisoner's Dilemma framework. Second, via simulations, we show that some more sophisticated algorithms exploit the seemingly-collusive ones. Following these results, we argue that the advancement of algorithms in sophistication and computational capabilities may, in some situations, provide a solution to the challenge of algorithmic seeming collusion, rather than exacerbate it.	2024	Ibrahim Abada, Xavier Lambin, Nikolay Tchakarov	Collusion by Mistake: Does Algorithmic Sophistication Drive Supra-Competitive Profits?
AI facilitated collusion	Pricing Algorithms as Collusive Devices	This paper undertakes a critical review of the prospect that self-learning pricing algorithms will lead to widespread collusion independently of the intervention and participation of humans. There is no concrete evidence, no example yet, and no antitrust case that self-learning pricing algorithms have colluded let alone increased the prospect of collusion across the economy.	2020	Cento Veljanovski	Pricing Algorithms as Collusive Devices
AI facilitated collusion	Pricing Algorithms: How Do They Work? When Do They Raise Concerns? How Should They Be Addressed?	Pricing algorithms have been an area of interest for competition authorities and international organizations in recent years. In 2023, the investigation initiated by the Turkish Competition Authority regarding e-commerce platforms implementing automatic pricing mechanisms has made this topic a significant agenda item in Turkish competition law practice. This study approaches the subject from both legal and economic perspectives, seeking answers to three fundamental questions: (1) What are pricing algorithms and how do they work? (2) Under what circumstances can they raise concerns regarding collusion? (3) When examining algorithm-related concerns, is there a need for change in the current competition law approach? The study first discusses current competition law practices by examining concerns about pricing algorithms in light of documents from competition authorities and international organizations. Then, to highlight potential competition law concerns, pricing algorithms are examined in two main categories: (1) "rule-based algorithms" and (2) "learning algorithms". Subsequently, considering that the most critical concern regarding algorithms is pricing collusion, the study addresses whether algorithms play an additional role in collusion. In this context, the potential effects of different types of algorithms on establishing and maintaining collusion are evaluated from a game theory perspective. Finally, the study questions whether there is a need for change in the current understanding of competition law in light of algorithm-related concerns.	2024	Emin Köksal, Bora Ikiler, Gediz Çınar	Pricing Algorithms: How Do They Work? When Do They Raise Concerns? How Should They Be Addressed? (Fiyatlandırma Algoritmaları: Nasıl Çalışır? Ne Zaman Endişe Yaratır? Nasıl Ele Alınmalıdır?)
AI facilitated collusion	Algorithmic Collusion of Pricing and Advertising on E-commerce Platforms	Online sellers have been adopting AI learning algorithms to automatically make product pricing and advertising decisions on e-commerce platforms. When sellers compete using such algorithms, one concern is that of tacit collusion-the algorithms learn to coordinate on higher than competitive prices which increase sellers' profits, but hurt consumers. This concern, however, was raised primarily when sellers use algorithms to decide on prices. We empirically investigate whether these concerns are valid when sellers make pricing and advertising decisions together, i.e., two-dimensional decisions. Our empirical strategy is to analyze competition with multi-agent reinforcement learning, which we calibrate to a large-scale dataset collected from Amazon.com products. Our first contribution is to find conditions under which learning algorithms can facilitate win-win-win outcomes that are beneficial for consumers, sellers, and even the platform, when consumers have high search costs. In these cases the algorithms learn to coordinate on prices that are lower than competitive prices. The intuition is that the algorithms learn to coordinate on lower advertising bids, which lower advertising costs, leading to lower prices for consumers and enlarging the demand on the platform. Our second contribution is an analysis of a large-scale, high-frequency keyword-product dataset for more than 2 million products on Amazon.com. Our estimates of consumer search costs show a wide range of costs for different product keywords. We generate an algorithm usage index based on the correlation patterns in prices and find a negative interaction between the estimated consumer search costs and the algorithm usage index, providing empirical evidence of beneficial collusion. We predict that in more than 50% of the product markets, consumers benefit from tacit collusion facilitated by algorithms. We also provide a proof that our results do not depend on the specific reinforcement learning algorithm that we analyzed. They would generalize to any learning algorithm that uses price and advertising bid exploration. Finally, we analyze the platform's strategic response through adjusting the ad auction reserve price or the sales commission rate. We find that reserve price adjustments will not increase profits for the platform, but commission adjustments will, while maintaining the beneficial outcomes for both sellers and consumers. Our analyses help alleviate some worries about the potentially harmful effects of competing learning algorithms, and can help sellers, platforms and policymakers to decide on whether to adopt or regulate such algorithms.	2025	Hangcheng Zhao, Ron Berman	Algorithmic Collusion of Pricing and Advertising on E-commerce Platforms
AI facilitated collusion AI and market power	Algorithmic Pricing, Market Outcomes, and Antitrust Concerns: Lessons from Recent Literature	The extensive use of algorithmic pricing, a type of pricing for products and services that is partially or fully based on computer codes, has attracted attention to its potential consequences on the overall economy-in particular, algorithmic pricing’s effects on market outcomes, such as price, volume of sales, or quality of services or products. Advances in computing capabilities, increases in the speed of internet communications, emergence of sophisticated computer software, and availability of public data have helped algorithmic pricing as a new and affordable option in firms’ decision-making toolboxes. Artificial intelligence (AI) pricing, which is at the heart of discussions about how firms are engaged in price setting, has been used more often as AI capabilities have developed rapidly in recent years. This article explores the effects of algorithmic pricing on market outcome. I start by discussing how algorithmic pricing can differ from traditional pricing in a market, in particular if there are few players in the market. I then discuss how the learning mechanisms of these algorithms play a role in market outcomes. I contrast asynchronous learning, where algorithms learn only from actual actions, with synchronous learning, which involves counterfactual analysis. Where an algorithm is on the spectrum between these learning approaches depends on factors such as algorithmic design, how much data is available, and managerial decision-making. I argue that firms’ reliance on algorithmic pricing should be viewed as a market outcome and the effect of algorithmic pricing can be assessed as a matter of profit maximization, as it is done in standard industrial organization models.	2025	Hassan Faghani	Algorithmic Pricing, Market Outcomes, and Antitrust Concerns: Lessons from Recent Literature
AI facilitated collusion Principles for AI and Competition Law Governance	Detecting Algorithmic Collusion	Imagine a consumer searching for a book on an online bookstore like Amazon and finding it priced at $30. Curious, she checks other online bookstores. Surprisingly, the price of the book is always the same—$30. Behind the scenes, these online bookstores might have used sophisticated algorithms enabling multiple parties to coordinate and agree on a single value, even in the presence of potential faults or discrepancies. This algorithm is known as a Byzantine agreement algorithm and can ensure that all retailers consistently display the same price, creating an appearance of uniformity across platforms. Collusion represents the greatest threat to competition. Although algorithmic collusion is in the spotlight of the Antitrust Division of the Department of Justice (DOJ) and the Federal Trade Commission (FTC), the risk of these algorithms largely escapes detection. Computers are finding new ways to collude by building algorithms that can ensure consistency and agreements in increasingly challenging situations. This paper examines algorithmic collusion, building a framework by extracting mechanisms from Byzantine agreement algorithms. This framework works conceptually; thus, it potentially detects collusion in a computer and non-computer setting. By extracting mechanisms from advanced algorithms, this paper proposes a computer science approach to complement the existing law-and-economics analytical tools used in antitrust. It offers new pathways to prosecute algorithmic collusion and advances antitrust enforcement in a digital economy. Antitrust does not need new rules but new analytical tools to enforce antitrust principles of competition and economic freedom in a digital computer run economy.	2025	Giovanna Massarotto	Detecting Algorithmic Collusion
AI facilitated collusion AI facilitated manipulation	Adversarial competition and collusion in algorithmic markets	Algorithms are now playing a central role in digital marketplaces, setting prices and automatically responding in real time to competitors’ behaviour. The deployment of automated pricing algorithms is scrutinized by economists and regulatory agencies, concerned about its impact on prices and competition. Existing research has so far been limited to cases where all firms use the same algorithm, suggesting that anti-competitive behaviour might spontaneously arise in that setting. Here we introduce and study a general anti-competitive mechanism, adversarial collusion, where one firm manipulates other sellers that use their own pricing algorithm. We propose a network-based framework to model the strategies of pricing algorithms on iterated two-firm and three-firm markets. In this framework, an attacker learns to endogenize competitors’ algorithms and then derive a strategy to artificially increase its profit at the expense of competitors. Facing a drastic loss of profits, competitors will eventually intervene and revise or turn off their pricing algorithm. To disincentivize this intervention, we show that the attacker can instead unilaterally increase both its profits and the profits of competitors. This leads to a collusive outcome with symmetric and supra-competitive profits, sustainable in the long run. Together, our findings highlight the need for policymakers and regulatory agencies to consider adversarial manipulations of algorithmic pricing, which might currently fall outside of the scope of current competition laws.	2023	Luc Rocher, J Tournier, Yves-Alexandre de Montjoye	Adversarial competition and collusion in algorithmic markets
AI facilitated collusion Principles for AI and Competition Law Governance	Algorithmic Pricing Facilitates Tacit Collusion: Evidence from E-Commerce	As the economy digitizes, menu costs fall, and firms can more easily monitor prices. These trends have led to the rise of automatic pricing tools. We employ a novel e-commerce dataset to examine the potential implications of these developments on price competition. We provide evidence from an RDD that the immediate impact of automatic pricing is a significant decline in prices. However, repricers have developed strategies to avoid the stark competitive realities of Bertrand-Nash competition. By employing plausibly exogenous variation in the execution of repricing strategies, we find that 'resetting' strategies (which regularly raise prices, e.g., at night) effectively coax competitors to raise their prices. While the resulting patterns of cycling prices are reminiscent of Maskin-Tirole's Edgeworth cycles, a model of equilibrium in delegated strategies fits the data better. This model suggests that if the available repricing technology remains fixed, cycling will increase, and prices could rise significantly in the future.	2022	Leon Musolff	Algorithmic Pricing Facilitates Tacit Collusion: Evidence from E-Commerce
AI facilitated collusion Principles for AI and Competition Law Governance	Algorithmic Pricing & Collusion; The Limits of Antitrust Enforcement	The combination of big data, large storage capacity, and computational power has strengthened the emergence of algorithms in making myriads of business decisions. It allows businesses to gain a competitive advantage by making automatic and optimize decision-making. In particular, the use of pricing algorithms allows businesses to match the demand and supply equilibrium by monitoring & setting dynamic pricing. It benefits consumers alike to see and act on fast-changing prices. However, on the downside, the widespread use of algorithms in an industry has the effect of altering the structural characteristics of the market such as price transparency, high-speed trading which increases the likelihood of collusion. The ability of the pricing algorithm to solve the cartel incentive problem by quickly detecting and punishing the deviant further strengthens the enforcement of the price-fixing agreement. In addition, the use of more advanced forms of algorithms such as self-learning algorithms allows businesses to achieve a tacitly collusive outcome in limited market characteristics even without communication between humans. This raises the fundamental challenge for anti-cartel enforcement as the current law in most jurisdictions is ill-equipped to deal with algorithmic facilitated tacit collusion. The legality of tacit collusion is questionable primarily because the pricing algorithm has the ability to alter the market characteristics where the tacitly collusive outcome is difficult to achieve; thus widening the scope of the so-called ‘oligopoly problem’. This paper studies the usage of pricing algorithms by businesses in online markets. In particular, the paper identifies the conditions under which the algorithm prices cause harm to consumers. It seeks to analyze how algorithms might facilitate or even cause the collusive outcome without human intervention. Further, it looks at the legal challenges faced by competition authorities around the globe to deal with algorithmic let collusion and examines the various approaches suggested to counteract it.	2024	Sumit Bhadauria, Lokesh Vyas	Algorithmic Pricing & Collusion; The Limits of Antitrust Enforcement
AI facilitated collusion	Pricing with algorithms	This paper studies Markov perfect equilibria in a repeated duopoly model where sellers choose algorithms. An algorithm is a mapping from the competitor's price to own price. Once set, the algorithms respond quickly. Customers arrive randomly and sellers can periodically revise their algorithms. The main results are that (i) for the simple two-price model with standard profit functions, monopoly pricing is the unique equilibrium outcome, and (ii) for general finite price grids, all equilibrium outcomes feature supra-competitive pricing. Sustenance of such collusion seems outside the scope of current antitrust laws for it does not involve any direct communication.	2025	Rohit Lamba, Sergey Zhuk	Pricing with algorithms
AI facilitated collusion	The risks of using algorithms in business: artificial price collusion	Increasingly, prices are set by algorithms rather than humans. Many competition authorities have voiced their concerns that this may enable firms (knowingly or otherwise) to avoid competitive pressure and collude. Exactly how would such algorithmic collusion work? And what can businesses and other organisations that use pricing algorithms expect from competition authorities in the future?	2020	Gareth Shier, Timo Klein	The risks of using algorithms in business: artificial price collusion
AI facilitated collusion Principles for AI and Competition Law Governance	Regulation of Algorithmic Collusion	Consider sellers in a competitive market that use algorithms to adapt their prices from data that they collect. In such a context it is plausible that algorithms could arrive at prices that are higher than the competitive prices and this may benefit sellers at the expense of consumers (i.e., the buyers in the market). This paper gives a definition of plausible algorithmic non-collusion for pricing algorithms. The definition allows a regulator to empirically audit algorithms by applying a statistical test to the data that they collect. Algorithms that are good, i.e., approximately optimize prices to market conditions, can be augmented to contain the data sufficient to pass the audit. Algorithms that have colluded on, e.g., supra-competitive prices cannot pass the audit. The definition allows sellers to possess useful side information that may be correlated with supply and demand and could affect the prices used by good algorithms. The paper provides an analysis of the statistical complexity of such an audit, i.e., how much data is sufficient for the test of non-collusion to be accurate.	2024	Jason D. Hartline, Sheng Long, Chenhao Zhang	Regulation of Algorithmic Collusion
AI facilitated collusion Principles for AI and Competition Law Governance	A (Mathematical) Definition of Algorithmic Collusion	Legal scholars and competition experts have expressed concerns regarding the ability of algorithms to collude without breaking competition law. This concern about algorithmic collusion has initiated a stream of literature that aims to construct algorithms that collude, but also a stream of literature that criticizes claims of algorithmic collusion. The debate partly originates from the fact that there exists no common definition of what it means for an algorithm to collude. In this paper, we contribute to the debate by proposing a mathematical definition of algorithmic collusion. Our definition allows for the systematic comparison of claims regarding algorithmic collusion, and thus paves the way for identifying algorithm design conducive to collusion. This is in turn necessary for the regulation and detection of algorithmic collusion in the future.	2024	Arnoud V. den Boer, Janusz M Meylahn	A (Mathematical) Definition of Algorithmic Collusion
AI facilitated collusion Principles for AI and Competition Law Governance	Algorithmic Collusion: Supra-Competitive Prices via Independent Algorithms	Motivated by their increasing prevalence, we study outcomes when competing sellers use machine learning algorithms to run real-time dynamic price experiments. These algorithms are often misspecified, ignoring the effect of factors outside their control, e.g. competitors' prices. We show that the long-run prices depend on the informational value (or signal to noise ratio) of price experiments: if low, the long-run prices are consistent with the static Nash equilibrium of the corresponding full information setting. However, if high, the long-run prices are supra-competitive---the full information joint-monopoly outcome is possible. We show this occurs via a novel channel: competitors' algorithms' prices end up running correlated experiments. Therefore, sellers' misspecified models overestimate own price sensitivity, resulting in higher prices. We discuss the implications on competition policy.	2020	Karsten Hansen, Kanishka Misra, Mallesh Pai	Algorithmic Collusion: Supra-Competitive Prices via Independent Algorithms
AI facilitated collusion	A comparison of naive and experienced bidders in common value offer auctions: A laboratory analysis	Laboratory economics experiments typically use financially motivated students as subjects. An ongoing issue is whether this is an appropriate subject pool since the students are typically inexperienced in the types of decision-making required of them in the lab. This paper addresses this issue in the context of common value offer auctions as we compare the -behaviour of experienced business executives in the construction contract industry ('experts') with that of ('naive') student subjects. Results of previous research of this sort have been equivocal; in some cases experts make the same errors as novices, in other cases they do not (Hogarth and Reder, I987). A series of sealed-bid, common value offer auctions in which bidders compete for the right to supply an item of unknown cost were conducted. Inherent to common value auctions (CVAs) is an adverse selection problem which may result in below normal or negative profits (the winner's curse). Experimental studies have documented the presence of the winner's curse with financially motivated student subjects in high price demand-side auctions (Kagel et al., I986; Kagel and Levin, I986). The experiments reported here generalise these earlier studies from bid to offer auctions. Also, in employing offer auctions we establish a setting with which our 'experts' are familiar, thus allowing their experience the best chance to manifest itself	1989	Douglas Dyer, John H Kagel, Dan Levin	A comparison of naive and experienced bidders in common value offer auctions: A laboratory analysis
AI facilitated collusion AI and Market Power	AI-Powered Trading, Algorithmic Collusion, and Price Efficiency	The integration of algorithmic trading with reinforcement learning, termed AI-powered trading, is transforming financial markets by reshaping how trading works. This study constructs a theoretical laboratory where financial markets function as information aggregation mechanisms, compelling investors to trade cautiously on private signals to preserve information rents. We find that informed AI speculators can autonomously sustain collusive supra-competitive profits without agreement, communication, or intent. AI collusion undermines competition and market efficiency, emerging robustly through two algorithmic mechanisms: (i) price-trigger strategies when information-insensitive investors are strongly prevalent and noise trading risk is low, and (ii) over-pruning bias in learning under other conditions.	2025	Winston Wei Dou, Itay Goldstein, Yan Ji	AI-Powered Trading, Algorithmic Collusion, and Price Efficiency
AI facilitated collusion AI and Market Power	Algorithmic Collusion with Imperfect Monitoring	We show that if they are allowed enough time to complete the learning, Q-learning algorithms can learn to collude in an environment with imperfect monitoring adapted from Green and Porter (1984), without having been instructed to do so, and without communicating with one another. Collusion is sustained by punishments that take the form of "price wars" triggered by the observation of low prices. The punishments have a finite duration, being harsher initially and then gradually fading away. Such punishments are triggered both by deviations and by adverse demand shocks	2021	Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò, Sergio Pastorello	Algorithmic Collusion with Imperfect Monitoring
AI facilitated collusion AI facilitated manipulations	Artificial Intelligence and Auction Design	Motivated by online advertising auctions, we study auction design in repeated auctions played by simple Artificial Intelligence algorithms (Q-learning). We find that first-price auctions with no additional feedback lead to tacit-collusive outcomes (bids lower than values), while second-price auctions do not. We show that the difference is driven by the incentive in first-price auctions to outbid opponents by just one bid increment. This facilitates re-coordination on low bids after a phase of experimentation. We also show that providing information about the lowest bid to win, as introduced by Google at the time of the switch to first-price auctions, increases the competitiveness of auctions.	2022	Martino Banchio, Andrzej Skrzypacz	Artificial Intelligence and Auction Design
AI facilitated collusion AI facilitated manipulations	Artificial Intelligence and Spontaneous Collusion	We develop a tractable model for studying strategic interactions between learning algorithms. We uncover a mechanism responsible for the emergence of algorithmic collusion. We observe that algorithms periodically coordinate on actions that are more profitable than static Nash equilibria. This novel collusive channel relies on an endogenous statistical linkage in the algorithms' estimates which we call spontaneous coupling. The model's parameters predict whether the statistical linkage will appear, and what market structures facilitate algorithmic collusion. We show that spontaneous coupling can sustain collusion in prices and market shares, complementing experimental findings in the literature. Finally, we apply our results to design algorithmic markets.	2022	Martino Banchio, Giacomo Mantegazza	Artificial Intelligence and Spontaneous Collusion
AI facilitated collusion	The impact of artificial intelligence design on pricing	The behavior of artificial intelligence (AI) algorithms is shaped by how they learn about their environment. We compare the prices generated by AIs that use different learning protocols when there is market interaction. Asynchronous learning occurs when the AI only learns about the return from the action it took. Synchronous learning occurs when the AI conducts counterfactuals to learn about the returns it would have earned had it taken an alternative action. The two lead to markedly different market prices. When future profits are not given positive weight by the AI, (perfect) synchronous updating leads to competitive pricing, while asynchronous can lead to pricing close to monopoly levels. We investigate how this result varies when either counterfactuals can only be calculated imperfectly and/or when the AI places a weight on future profits. Lastly, we investigate performance differences between offline and online play.	2023	John Asker, Chaim Fershtman, Ariel Pakes	The impact of artificial intelligence design on pricing
AI facilitated collusion	Algorithmic Collusion by Large Language Models	The rise of algorithmic pricing raises concerns of algorithmic collusion. We conduct experiments with algorithmic pricing agents based on Large Language Models (LLMs). We find that (1) LLM-based agents are adept at pricing tasks, (2) LLM-based pricing agents quickly and autonomously reach supracompetitive prices and profits in oligopoly settings, and (3) variation in seemingly innocuous phrases in LLM instructions (“prompts”) may substantially influence the degree of supracompetitive pricing. Off-path analysis using novel techniques uncovers price-war concerns as contributing to these phenomena. Our results extend to auction settings. Our findings uncover unique challenges to any future regulation of LLM-based pricing agents, and generative AI pricing agents more broadly	2025	Sara Fish, Yannai A Gonczarowski, Ran Shorrer	Algorithmic Collusion by Large Language Models
AI facilitated collusion	The Effect of Explicit Communication on pricing: Evidence from the Collapse of a Gasoline Cartel	We study the collapse of collusion in Québec's retail gasoline market following a Competition Bureau investigation, and show that it involved two empirical regularities: high margins, and asymmetric price adjustments. Using weekly, station-level prices we test whether collusion was successful, and whether asymmetric adjustments were part of the cartel's strategy. We do so in the markets targeted by the investigation, and in markets throughout the province with similar pre-collapse pricing (cyclical markets). Our results suggest that stations in both target and cyclical markets adjusted pricing following the announcement: margins fell (by 30%/15% in target/cyclical markets), and adjustments became more symmetric.	2014	Robert Clark, Jean-François Houde	The Effect of Explicit Communication on pricing: Evidence from the Collapse of a Gasoline Cartel
AI facilitated collusion	Learning to Coordinate: A Study in Retail Gasoline	This paper studies equilibrium selection in the retail gasoline industry. We exploit a unique dataset that contains the universe of station-level prices for an urban market for 15 years, and that encompasses a coordinated equilibrium transition mid-sample. We uncover a gradual, three-year equilibrium transition, whereby dominant firms use price leadership and price experiments to create focal points that coordinate market prices, soften price competition, and enhance retail margins. Our results inform the theory of collusion, with particular relevance to the initiation of collusion and equilibrium selection. We also highlight new insights into merger policy and collusion detection strategies.	2018	David P. Byrne, Nicolas de Roos	Learning to Coordinate: A Study in Retail Gasoline
AI facilitated collusion	Collusion with Asymmetric Retailers: Evidence from a Gasoline Price-Fixing Case	We point out a fundamental diffi culty of successfully colluding in retail markets with heterogeneous fi rms, and characterize the mechanism recent gasoline cartels in Canada used to sustain collusion. Heterogeneity in cost and network size necessitates arrangements whereby participants split the market unequally to favor stronger players. We characterize empirically the strategy and transfer mechanism using court documents containing summaries and extracts of conversations between participants. The mechanism implements transfers based on adjustment delays during price changes. We estimate that these delays can translate into substantial transfers and provide examples in which they can substantially reduce deviation frequency.	2013	Robert Clark and Jean-François Houde	Collusion with Asymmetric Retailers: Evidence from a Gasoline Price-Fixing Case
AI facilitated collusion	An empirical analysis of algorithmic pricing on Amazon marketplace.	The rise of e-commerce has unlocked practical applications for algorithmic pricing (also called dynamic pricing algorithms), where sellers set prices using computer algorithms. Travel websites and large, well known e-retailers have already adopted algorithmic pricing strategies, but the tools and techniques are now available to small-scale sellers as well. While algorithmic pricing can make merchants more competitive, it also creates new challenges. Examples have emerged of cases where competing pieces of algorithmic pricing software interacted in unexpected ways and produced unpredictable prices [37], as well as cases where algorithms were intentionally designed to implement price fixing [5]. Unfortunately, the public currently lack comprehensive knowledge about the prevalence and behavior of algorithmic pricing algorithms in-the-wild. In this study, we develop a methodology for detecting algorithmic pricing, and use it empirically to analyze their prevalence and behavior on Amazon Marketplace. We gather four months of data covering all merchants selling any of 1,641 best-seller products. Using this dataset, we are able to uncover the algorithmic pricing strategies adopted by over 500 sellers. We explore the characteristics of these sellers and characterize the impact of these strategies on the dynamics of the marketplace.	2016	Le Chen, Alan Mislove, Christo Wilson	An empirical analysis of algorithmic pricing on Amazon marketplace
AI facilitated collusion	“Why Do Gas Station Prices Constantly Change? Blame the Algorithms”	Retailers are using artificial-intelligence software to set optimal prices, testing textbook theories of competition; antitrust officials worry such systems raise prices for consumers	2017	The Wall Street Journal, Sam Schechner	“Why Do Gas Station Prices Constantly Change? Blame the Algorithms”
AI facilitated collusion	Algorithms and Antitrust: A Framework with Special Emphasis on Coordinated Pricing	The debate about algorithmic collusion has solidified to a state where agencies like the EU-Commission or the UK CMA have acknowledged its relevance for the cartel prohibition in their latest Horizontal Guidelines. In addition, national legislators in Germany and Italy have, in the last months, enacted special antitrust provisions with the intention, among other things, to tackle algorithmic collusion even outside the scope of the cartel prohibition. We argue that established legal principles behind the definition of cartel conduct are challenged by the means and forms of how algorithms can impact pricing. We put forward that key for any case analysis under the cartel prohibition and the new type of legislation is a counterfactual assessment, which reflects the capabilities of artificial intelligence-based pricing technology. Such counterfactual assessment hinges on the type of pricing algorithms and the effects that the blocking of certain functions of algorithmic pricing would have on consumer welfare. In essence, the counterfactual should not be “algorithmic pricing without collusion.” Rather, the counterfactual must be shaped around a type of modified pricing mechanism, which the agencies or courts need to define in the first place. Only then is it possible to assess if modified types of pricing will increase consumer rent or not. We develop a taxonomy of cases for the cartel prohibition, and we describe paradigms for the development of remedies within the realm of the new legislation.	2025	Roman Inderst, Stefan Thomas	Algorithms and Antitrust: A Framework with Special Emphasis on Coordinated Pricing
AI facilitated collusion	De-Humanizing Antitrust: The Rise of the Machines and the Regulation of Competition	Increasingly, firms are knitting together newly available mass data collection, Internet-driven interconnective power, and automated algorithmic selling with their traditional supply-chain and sales functions. Traditional sales functions such as competitive intelligence gathering and pricing are being delegated to software “robo-sellers.” This Article offers the first descriptive and normative study of the implications of this shift away from humans to machines (the “robo-sellers”) for antitrust law. This change is a critical challenge for antitrust law – both in how it is currently applied and in highlighting and exacerbating its existing weaknesses. First – and critically – robo-sellers will increase the risk that oligopolists will coordinate prices above the competitive level, thereby harming consumers. The Sherman Act contains a well-known gap in its coverage under which oligopolists that achieve price coordination interdependently, without communication or facilitating practices, generally escape antitrust enforcement, even when their actions yield supracompetitive pricing that harms consumers. Because robo-sellers possess traits that will make them better than humans at achieving supracompetitive pricing without communication, all things being equal, they will increase consumer harm due to this gap. A second problem concerns blackletter antitrust law in dealing with price coordination through communication or facilitating practices; current doctrine requires that there be an anticompetitive “agreement” for there to be a violation of the Sherman Act for price fixing. Under standard models, even where oligopolists have independent incentives to price supracompetitively, they can often do better via an agreement; moreover, in other cases, competing firms can only achieve supracompetitive pricing by explicit collusion. In these cases, usually analyzed as a prisoner’s dilemma in which the Nash equilibrium is to “cheat” on the cartel, an agreement is required to avoid the inferior (from the price-fixers’ perspective) outcome. In order to find such an “agreement,” courts, government enforcers, and practitioners tend to focus on finding “intent,” efforts to sowing fear and distrust, and discovering a “meeting of the minds.” These standard inquiries derive from a more than a century-old embedded assumption that antitrust regulates sales by human actors; they will be a poor fit in addressing robo-sellers, which will function differently and which will likely not create the same kinds of evidence that these inquiries rely on. What can be done about the anticompetitive effects of robo-selling? This Article assesses several possible solutions, but find that they will be quite difficult to reconcile with current antitrust law. It conclude that, at least as a feasible second-best result, incorporating an evolving approach to robo-sellers may be a worthwhile expansion of the FTC’s ongoing regulatory program that has already begun target the competition and consumer protection aspects of consumer data collection by sellers. For example, the FTC has already begun to consider the effects of mass data collection and algorithmic processing on consumers from the perspective of disclosure and discrimination (both price and social); efficiencies should exist in broadening the inquiry to include effects on price coordination and cartel behavior.	2014	Salil K Mehra	De-Humanizing Antitrust: The Rise of the Machines and the Regulation of Competition’
AI facilitated collusion Principles for AI regulation and Competition law	Competition policy in the digital age	The combination of big data with technologically advanced tools, such as pricing algorithms, is increasingly diffused in everyone’s life today, and this is changing the competitive landscape in which many companies operate and the way in which they make commercial and strategic decisions. While the size of this phenomenon is to a large extent unknown, a growing number of firms are using computer algorithms to improve their pricing models, customise services and predict market trends. This phenomenon is undoubtedly associated to significant efficiencies, which benefit firms as well as consumers in terms of new, better and more tailored products and services. However, a widespread use of algorithms has also raised concerns of possible anticompetitive behaviour as they can make it easier for firms to achieve and sustain collusion without any formal agreement or human interaction. This paper focuses on the question of whether algorithms can make tacit collusion easier not only in oligopolistic markets, but also in markets which do not manifest the structural features that are usually associated with the risk of collusion. This paper discusses some of the challenges algorithms present for both competition law enforcement and market regulation. In particular, the paper addresses the question of whether antitrust agencies should revise the traditional concepts of agreement and tacit collusion for antitrust purposes, and discusses how traditional antitrust tools might be used to tackle some forms of algorithmic collusion. Recognising the multiple risks of algorithms and machine learning for society, the paper also raises the question of whether there is need to regulate algorithms and the possible consequences that such a policy choice may have on competition and innovation	2017	OECD	Algorithms and Collusion: Competition Policy in the Digital Age’
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms and Collusion – Note from the European Union	-	2017	OECD	Algorithms and Collusion – Note from the European Union
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms and Collusion – Background note by the Secretariat’	The combination of big data with technologically advanced tools, such as pricing algorithms, is increasingly diffused in today everyone’s life, and it is changing the competitive landscape in which many companies operate and the way in which they make commercial and strategic decisions. While the size of this phenomenon is to a large extent unknown, there are a growing number of firms using computer algorithms to improve their pricing models, customise services and predict market trends. This phenomenon is undoubtedly associated to important efficiencies, which benefit firms as well as consumers in terms of new, better and more tailored products and services. However, a widespread use of algorithms has also raised concerns of possible anticompetitive behaviour as they can make it easier for firms to achieve and sustain collusion without any formal agreement or human interaction. In particular, this paper focuses on the question of whether algorithms can make tacit collusion easier not only in oligopolistic markets, but also in markets which do not manifest the structural features that are usually associated with the risk of collusion. This background note discusses some of the challenges of algorithms for both competition law enforcement and market regulation. In particular, the paper addresses the question of whether antitrust agencies should revise the traditional concepts of agreement and tacit collusion for antitrust purposes, and discusses how traditional antitrust tools might be used to tackle some forms of algorithmic collusion. Recognising the multiple risks of algorithms and machine learning for society, the paper also raises the question of whether there is need to regulate algorithms and the possible consequences that such a policy choice may have on competition and innovation.	2017	OECD	Algorithms and Collusion – Background note by the Secretariat’
AI facilitated collusion Principles for AI Regulation and Competition Law AI and Market Power	OECD Workshop Addresses Algorithms and Collusion Issues	On June 21-23, the OECD held a roundtable on the theme of “Algorithms and Collusion,” as part of a wider work stream on competition in the digital economy. The OECD roundtable reflects a shift in the debate over the antitrust implications of big data from concerns about the potential for companies to hoard big data, creating barriers to entry and market power, to concerns about companies with access to the same or comparable big data using algorithms to collude. The OECD’s background paper (the OECD note) and the other papers prepared for the roundtable constitute the most authoritative discussion of algorithms and collusion to date. These materials discuss the way in which algorithms may change market structures and behavior in ways not contemplated by traditional antitrust thinking, how antitrust authorities can address these issues and a range of potential regulatory responses. The OECD’s proposals are wide-ranging and potentially very significant; for instance, expanding the concept of “agreement” subject to antitrust rules, searching out anti-competitive conduct in new markets, expanding merger review especially in relation to coordinated or conglomerate effects and considering new regulatory initiatives. So far, the individual authorities participating in the roundtable seem unpersuaded of the need for dramatic reforms of the type proposed by the OECD. The EU background paper comments that “pricing algorithms can be analysed by reference to the traditional reasoning and categories used in EU competition law” (para. 35). On the other hand, these submissions point to a number of specific practices not discussed in detail in the OECD note, and all participating authorities stressed the importance of algorithm issues and the need for further study. This briefing summarizes the issues and proposals discussed at the roundtable, following the structure of the OECD note for convenience. Different perspectives or comments by the individual jurisdictions participating in the roundtable are also discussed below.	2017	Norton Rose Fulbright	OECD Workshop Addresses Algorithms and Collusion Issues
AI facilitated collusion AI and Market Power	Robo-Seller and Prosecutions and Antitrust’s Error-Cost Framework	Over the past decade, we have seen the spread of software algorithms and automated trading beyond their initial economic beachhead in relatively software-friendly areas such as Internet searches and financial markets. As recently as the middle of the last decade, it was considered unlikely that driverless vehicles such as the Google/Waymo car plying the roads of California would be possible anytime soon, since software was only fit for “highly structured situations.”2 Obviously, times have changed quickly.	2017	Salil K Mehra	Robo-Seller and Prosecutions and Antitrust’s Error-Cost Framework
AI facilitated collusion	Rise of the Machines: Emerging Antitrust Issues Relating to Algorithm Bias and Automation	Artificial intelligence-driven automation has been hailed as a key driver of the “Fourth Industrial Revolution,” but there have also been warnings that they “have the potential to disrupt the current livelihoods of millions of Americans.” As automation reshapes competitive landscapes across industries, new antitrust issues will undoubtedly arise. This paper focuses on two areas of automation that have begun to draw attention from antitrust regulators: algorithm bias in online services and potential price fixing resulting from automated pricing algorithms.	2017	Sheng Li, Claire (Chunying) Xie	Rise of the Machines: Emerging Antitrust Issues Relating to Algorithm Bias and Automation
AI facilitated collusion	Algorithmic-facilitated coordination: Market and Legal Solutions	Technological developments, it was hoped, would bring about more competition. The ability to connect faster and more easily with numerous suppliers on-line through digital platforms, as well as the use of algorithms by consumers in order to compare more offers in a more efficient and sophisticated manner, strengthened pressures on suppliers to provide better and cheaper products and services. These advantages, however, are currently threatened by algorithmic-facilitated coordination. Algorithms make coordination – both implicit or tacit – much easier and quicker than ever before. Such coordination may bring about many positive effects. For example, they enable suppliers to better coordinate their conduct with the demands of consumers, thereby saving scarce resources, and responding much faster to demand trends. At the same time, and based on similar technological abilities, algorithms ease coordination among competing suppliers. Indeed, coordination no longer requires firms to operate in oligopolistic markets; and firms can more quickly and easily detect and punish deviations from the status-quo, thereby reducing incentives for shirking. As our assumptions about which market conditions must exist for firms to coordinate are altered, the number of red flags that are raised across industries rises. As Ezrachi and Stucke write, this is the end of competition as we know it. This requires us to explore which tools – either market-based or regulatory – can be used, if at all, in order to reduce the negative welfare effects of algorithmic coordination among competitors. Given that some of the assumptions that stand at the basis of the current rule under which tacit collusion is not considered an “agreement in restraint of trade” do not hold anymore, it is time to determine whether our laws are fit to deal with the digitized world; whether we are looking under the lamp while most of the occurrence in the real world is happening outside its scope of light. In other words, can we widen the scope of the light by simply using a stronger light bulb in the same lamp, or do we need to create a new source of light altogether? Accordingly, this short note focuses on three issues that arise from this technological challenge. First, it explores the effects of algorithms on the ability of suppliers to coordinate their conduct. Second, it explores the ability of existing technological and regulatory tools to deal effectively with algorithmic-facilitated coordination. The final part briefly explores the promises as well as the limits of market solutions to welfare-reducing algorithmic coordination, which can be complementary or provide at least some viable alternative for the possible failure of regulation to deal with algorithmic-facilitated coordination. Issues of vertical integration and coordination, while important, are not addressed in this note	2017	Michal S Gal	Algorithmic-facilitated coordination: Market and Legal Solutions
AI facilitated collusion AI and Market Power	How Pricing Bots Could Form Cartels and Make Things More Expensive	How competitive is our market economy? Not as much as it ought to be. And the growth of big data threatens to make things even worse. Antitrust regulators already struggle to keep markets competitive. How will they fare in a world where intelligent pricing algorithms subtly collude with one another?	2016	Maurice E Stucke, Ariel Ezrachi	How Pricing Bots Could Form Cartels and Make Things More Expensive
AI facilitated collusion Principles for AI Regulation and Competition Law	Two Artificial Neural Networks Meet in an Online Hub and Change the Future (Of Competition, Market Dynamics and Society)	In the future, one may imagine a new breed of antitrust humor. Jokes might start along the following lines: “Two Artificial Neural Network and one Nash equilibrium meet in an online (pub) hub. After a few milliseconds, a unique silent friendship is formed…” Back to the present; we are not sure how this joke might end. Nor can we estimate how funny future consumers would find it. We can, however, explain, at present, how technological advancements have changed, and will continue to change, the dynamics of competition and subsequently the distribution of wealth in society. How algorithms may be used in stealth mode to stabilize and dampen market competition while retaining the façade of a competitive environment. That tale is at the heart of this paper. We first raised algorithmic tacit collusion in 2015. Our recent book, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy, provides further context and analysis. We illustrate how online tacit collusion may emerge when products are generally homogeneous, sellers do not benefit from brand recognition or loyalty, and markets are transparent and concentrated. Since our book elaborates on the four collusion scenarios, we begin here by outlining one model of tacit collusion and its manifestation online. Taking note of advancements in technology and emerging policies, we move the debate forward in reviewing the possible harm and means to address it. We illustrate with several case studies how the move to an online pricing environment, under certain market conditions, may harm the buyers’ welfare. We note how new technologies may undermine enforcers’ attempts to intervene - as stealth becomes a feature of future strategies. That tale, of course, is not immune from disruptive strategies. We consider the testing of counter-measures in an “algorithmic collusion incubator” to better understand what effectively destabilizes algorithmic tacit collusion. Further, we consider the effects and likelihood of secret dealings. We note how, somewhat counter-intuitively, secret deals in an online environment could reduce, at times, our welfare	2017	Maurice E Stucke, Ariel Ezrachi	Two Artificial Neural Networks Meet in an Online Hub and Change the Future (Of Competition, Market Dynamics and Society)
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms and Competition: Friends or Foes?.	There has been a recent wave of distress in some corners of the antitrust community about the hypothesis that, with the fast development of machine learning and the growing popularity of pricing algorithms, firms are developing new sophisticated strategies to collude under the radar of antitrust watchdogs. Among several possible theories of harm, there is the concern that artificially intelligent machines may be causing competitive harm by coordinating prices in a much more efficient way than what a human being could ever aspire to do. We will focus this paper on this particular risk and leave aside other possible theories of harm. The lively debate that is growing around this topic, while initially kept more at an academic level, is now reaching antitrust practitioners, competition authorities, governments and international fora such as the OECD, which hosted in June 2017 a Roundtable on Algorithms and Collusion. What’s the verdict? Well, the jury is still out and while some commentators remain skeptical about the risks of algorithms, possibly concerned with the burden that a stronger antitrust enforcement could pose on businesses, others claim that this is a sensational discussion not to be taken more seriously than stories about machines taking control over humans. So which of these visions is true? Are competition policy debates turning into arenas to discuss science fiction stories that serve only to stimulate our intellect? Or do we increasingly live in a world where some market players, using complex computer codes, can exploit and harm those who do not dominate technology? While it is probably exaggerated to say that computer algorithms will dramatically change everything we know about competition, it would also be unwise to ignore the clear signals that markets are changing, as well as the resulting implications for competition policy. Apart from the evidence of cartels in multiple jurisdictions that have been facilitated by advanced technology, the fact that the risks of algorithms are being taken very seriously by heads of agency and even heads of State suggests, at the very least, that this problem deserves deeper scrutiny before one chooses to disregard it. In the analysis of this problem, it is equally important to avoid over simplistic approaches, such as assuming a priori that algorithms always facilitate anti-competitive agreements prohibited by competition law. Although in some cases this might be true, in other cases algorithms may raise new questions that require a more articulated analysis. In particular, it is important to evaluate whether algorithms might allow companies not only to collude in a wider spectrum of market structures, but also to do so without necessarily triggering a violation of competition laws, challenging thus existing antitrust approaches.	2017	Antonio Capobianco, Pedro Gonzaga	Algorithms and Competition: Friends or Foes?
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms, Artificial Intelligence, and Joint Conduct’.	The ability of algorithms and artificial intelligence to monitor and set prices is increasing in sophistication, effectiveness and independence from human involvement at an exponential rate. The growth in this area, which is seen simultaneously across a range of AI applications, is such that no one — even its creators — is likely to fully appreciate AI’s capabilities until sometime after they have been realized. Pricing “bots” are already capable of engaging in behavior that we would not hesitate to call “parallel conduct” if it were performed by humans, and they will only get better at it. Indeed, the day may not be so far off when the pricing bot of one firm is fully capable of colluding — in every meaningful sense — with the pricing bot of a competing firm. At that point, we may have “conspiracy” cases under Section 1 of the Sherman Act that look very much like the cases we have today, except that the parts now played by humans are played by robots.2 The few existing antitrust cases involving pricing algorithms have not crossed this Rubicon, or really even approached it. They do not involve joint conduct by bots, in any sense. Instead, these cases involve human beings reaching familiar price-fixing agreements and then implementing them algorithmically. While these cases may create special problems of detection and proof, at least for the moment they do not seem to require any shift in the conceptual apparatus we use to solve antitrust problems. There is reason to think such a shift may be coming, however. Joint conduct by robots is likely to be different — harder to detect, more effective, more stable and persistent — than traditional joint conduct by humans. For example, one of the basic precepts of the Sherman Act is that “unilateral” conduct by firms in the same market is not unlawful under Section 1, even if the conduct is closely interdependent and predictably yields supracompetitive prices that would be per se unlawful if achieved by agreement. An unspoken premise of this time-honored rule is that such interdependent conduct is likely to be relatively unstable in the absence of an agreement, and therefore, with any luck, the supracompetitive effects generally will be shorter lived and less pernicious than if they were achieved through true joint conduct. But this premise may have less force in a world of bots, who can interpret and respond to the actions of their competitors with far more precision, agility and consistency than their human counterparts. By simply allowing these bots to go to work, it is easy to imagine an effectively permanent pricing stasis settling over many markets, and not always with procompetitive effects. How will enforcers approach such conduct, much less disrupt or prevent it? What duties should we impose on human beings to ensure their bots behave, and what culpability should they have when their bots go astray? The next ten years will begin to provide the answers, but the technology is already well ahead of the law, and the growing pains are likely to be immense.	2017	Dylan I Ballard, Amar S Naik	Algorithms, Artificial Intelligence, and Joint Conduct’
AI facilitated collusion	Algorithmic Collusion: Problems and CounterMeasures	In this paper, we explore how technological advancements have changed, and will continue to change, the dynamics of competition and subsequently the distribution of wealth in society. How algorithms may be used in stealth mode to stabilize and dampen market competition while retaining the façade of competitive environment. We first raised algorithmic tacit collusion in 2015.1 In 2016 we provided further context and analysis in our book, Virtual Competition: The Promise and Perils of the AlgorithmDriven Economy. 2 We illustrated how online tacit collusion may emerge when products are generally homogeneous and sellers do not benefit from brand recognition or loyalty, and when markets are transparent and concentrated.	2017	OECD	‘Algorithmic Collusion: Problems and CounterMeasures
AI facilitated collusion	Antitrust in digital markets in the EU: policing price bots	A number of authors have in recent years stated that current antitrust rules may not be able to police supra-competitive price levels (or indeed other undesirable market outcomes) which may result from the use of price robots. This paper discusses what tools are available in EU antitrust law to tackle collusion by price bots, based on the existing legislation, the case law of the European courts and the practice of the European Commission. It argues that unlawful collusion can result from the disclosure of sensitive information from one undertaking to another, even in the absence of anticompetitive intent. On this basis, even self-learning pricing algorithms could be caught by the prohibition of Article 101 TFEU. Furthermore, undertakings can be liable for the actions of the (self-learning) algorithms they create or use. Undertakings have a positive obligation to ensure compliance with the EU antitrust rules and cannot plead ignorance of what their employees or price bots are doing. And even if there would be circumstances where undertakings could not be found to have been negligent in how they supervise their employees and price bots, the toolbox of the European Commission is large enough to stop practices for which no undertaking is to blame.	2017	Jan Blockx	Antitrust in digital markets in the EU: policing price bots’
AI facilitated collusion Principles of AI Regulation and Competition Law	Algorithmic Tacit Collusion: A Regulatory Approach	In light of the ongoing debate on algorithmic collusion, this paper intends to answer the following question: should algorithmic tacit collusion be prohibited under EU competition law? By algorithmic tacit collusion it is meant the capability of algorithmic pricing agents to unilaterally engage into tacitly collusive strategies without human intervention (we also call it ‘machine-to-machine cooperation’ or ‘algorithmic interdependent pricing’). Essentially, this practice raises the very same issues as the well-known oligopoly problem. Therefore, to make it prohibited one could envisage the traditional proposed solution of disentangling the categories of article 101 TFEU from the notions of an ‘act of reciprocal communication between firms’ or ‘meeting of minds’. The paper sets out to discuss this option and its implications. It argues that, from a competition law standpoint, although algorithmic tacit collusion remains undesirable, the notions of agreement and concerted practices should not be changed to encompass it. Rather, it embraces a regulatory perspective referred to as ‘algorithms by design’ which relies on introducing a legal obligation for firms to program algorithms in such a way as to prevent them from setting oligopolistic prices. In particular, while exploring this regulatory proposal, the paper discusses the peculiar case of algorithms that, though designed not to violate antitrust law, end up charging collusive prices. In this regard, the paper develops a second proposal: it introduces the idea of ‘outcome visibility’ to nail firms to their responsibility. This concept implies the idea that even if firms are not aware that their pricing algorithms are implementing a collusive strategy, they cannot ignore their visible market outcome.	2023	Valeria Caforio	Algorithmic Tacit Collusion: A Regulatory Approach
AI facilitated collusion Principles of AI Regulation and Competition Law	Algorithms and Competition Law - Status and Challenges	Algorithmic collusion has been looming large in the competition law discourse, driven by some case law and pioneering publications from the OECD, Gal, Ezrachi/Stucke, and others. Recently, though, other topics have pushed the implications of algorithmic market conduct for competition a little aside. This stands in stark contrast to the rapid spread of algorithmic software as a key tool for doing business, not only on digital gatekeepers’ platforms but also in the brick-and-mortar activities of small and medium enterprises (SMEs). Correspondingly, algorithms remain a recurring topic for competition authorities around the globe. As to their technology, algorithmic tools are developing ever faster, from relatively “static” towards autonomous, machine-learning systems that learn from data and deduce their action parameters with a higher degree of automation. The employment of deep learning artificial neural networks, in particular, has become a business reality. In view of this divergence between practical relevance and academic attention, the Center for Intellectual Property and Competition Law (CIPCO) at Zurich University undertook – in cooperation with the Academic Society for Competition Law (Ascola) and the Swiss Competition Commission (ComCo) – an interdisciplinary research project (the “Research Project”) on algorithmic market activity and competition (law). Based on the project results, this paper takes stock of the discourse regarding algorithmic market activity and competition law and economics. Furthermore, it discusses three topics which the Research Project has indicated to be of particular importance for competition law’s application to algorithmic market activity, namely rules on mitigating algorithmic collusion risks that are viable for main street businesses, unilateral conduct employing generative “artificial intelligence”, and the control of algorithmic compliance through audits.	2024	Peter Georg Picht, Anna-Katharina Leitz	Algorithms and Competition Law - Status and Challenges
AI facilitated collusion Principles of AI Regulation and Competition Law	A Few Reflections on the Recent Caselaw on Algorithmic Collusion	The article discusses the limited number of cases where the use of algorithms was found to be a factor that contributed to collusion. We discuss the difficulties that competition agencies face in addressing such conducts, with particular focus where there is no direct communication between competitors but collusion is derived through the use of a third party algorithm. We offer some reflections on the current enforcement approach and we argue that with the current development in enforcement, there is a need for the competition authority to show some type of communication between the members of the cartel in order to substantiate the anticompetitive conduct.	2020	Ioannis Kokkoris	A Few Reflections on the Recent Caselaw on Algorithmic Collusion
AI facilitated collusion Principles of AI Regulation and Competition Law	Algorithmic Collusion: Where Are We and Where Should We Be Going?	Research on the possibility of algorithmic collusion has rapidly expanded in recent years and has come to the attention of competition authorities worldwide. Claims regarding the ability of pricing algorithms to collude have, however, been studied through the lens of classical results pertaining to human collusion by assessing the emergence of reward-punishment schemes to sustain high prices. In this article, we argue that this is neither necessary nor sufficient to cause concern for policymakers. Leveraging results of recent research, we propose criteria to define classes of learning algorithms of special concern, as they are prone to set high prices in a robust and persistent way.	2025	Ibrahim Abada, Joseph E Harrington Jr, Xavier Lambin, Janusz M Meylahn	Algorithmic Collusion: Where Are We and Where Should We Be Going?
AI facilitated collusion	Algorithmic Collusion: Insights from Deep Learning	Increasingly, firms use algorithms powered by artificial intelligence to set prices. Previous research simulated interactions among Q-learning algorithms in an oligopoly model of price competition. The algorithms learn collusive strategies but require a long time that corresponds to several years to do so. We show that pricing algorithms using deep learning (DQN) can collude significantly faster. The availability of these more powerful pricing algorithms enables simulations in larger markets. Collusion disappears in wide oligopolies with up to 10 firms. However, incorporating knowledge of the learning behavior by reformulating the state representation increases the ability to collude effectively.	2021	Matthias Hettich	Algorithmic Collusion: Insights from Deep Learning
AI facilitated collusion AI mergers and cooperation	The Role of Secondary Algorithmic Tacit Collusion in Achieving Market Alignment	The antitrust risks associated with the use of the same hub’s pricing algorithm by many sellers are now well-accepted. But what if many rivals use several different hubs for dynamic pricing? The common assumption is that in such instances, competition among the pricing hubs would support competition among the sellers. However, in this paper we argue differently and introduce the concept of secondary algorithmic tacit collusion, which leads to anticompetitive effects, independent of the conditions on the primary market. This phenomenon may lead to the evils of price-fixing but on far a wider scale. Contrary to traditional tacit collusion, this aggregated form of collusion, through the use of algorithmic hub-and-spoke structures, can occur in markets with many competitors and with seemingly competitive dynamics. We outline how the combination of hub-and-spoke frameworks on the primary market and conscious parallelism on the secondary market for algorithmic pricing services can lead to secondary tacit collusion. Addressing its anticompetitive effects requires competition agencies to consider the interaction between price setters in the secondary markets, while taking note of the hub-and-spoke structures on the primary market.	2023	Ariel Ezrachi, Maurice E Stucke	The Role of Secondary Algorithmic Tacit Collusion in Achieving Market Alignment
AI facilitated collusion	Algorithmic and Human Collusion	I study self-learning pricing algorithms and show that they are collusive in market simulations. To derive a counterfactual that resembles traditional tacit collusion, I conduct market experiments with humans in the same environment. Across different treatments, I vary the market size and the number of firms that use a pricing algorithm. I demonstrate that oligopoly markets can become more collusive if algorithms make pricing decisions instead of humans. In two-firm markets, prices are weakly increasing in the number of algorithms in the market. In three-firm markets, algorithms weaken competition if most firms use an algorithm and human sellers are inexperienced.	2024	Tobias Werner	Algorithmic and Human Collusion
AI facilitated collusion AI exclusion and exploitation	Competition (Law) in the Era of Algorithms	Algorithm-driven computer programs have become key instruments for market success in a digitalized economy. They can generate positive effects on consumer welfare and welfare in general. On the other hand, algorithms may foster tacit collusion, adversely affect consumer choice, even pose a threat to pluralism. Especially since algo-driven market interactions call traditional economic models into question, it is still unclear whether and how the new challenges can be addressed within the existing framework of (competition) law or whether new legal tools, such as algorithm-focused regulation, must be developed. To approach these questions, the Center for Intellectual Property and Competition Law (CIPCO) at the University of Zurich held a workshop in February 2018. The first part of the workshop focused on technical and economic fundamentals, the second on effects on consumers, and the third part on the existing case-law, as well as on the practice and policy of competition agencies. The present paper reflects the discussions and results of the workshop.	2018	Peter Georg Picht, Benedikt Freund	Competition (Law) in the Era of Algorithms
AI facilitated collusion Principles for AI Regulation and Competition Law	Antitrust and the Robo-Seller: Competition in the Time of Algorithms	Increasingly, firms are knitting together newly available mass data collection, Internet-driven interconnective power, and automated algorithmic selling with their traditional supply-chain and sales functions. Traditional sales functions such as competitive intelligence gathering and pricing are being delegated to software “robo-sellers.” This Article offers the first descriptive and normative study of the implications of this shift away from humans to machines (the “robo-sellers”) for antitrust law. This change is a critical challenge for antitrust law – both in how it is currently applied and in highlighting and exacerbating its existing weaknesses. First – and critically – robo-sellers will increase the risk that oligopolists will coordinate prices above the competitive level, thereby harming consumers. The Sherman Act contains a well-known gap in its coverage under which oligopolists that achieve price coordination interdependently, without communication or facilitating practices, generally escape antitrust enforcement, even when their actions yield supracompetitive pricing that harms consumers. Because robo-sellers possess traits that will make them better than humans at achieving supracompetitive pricing without communication, all things being equal, they will increase consumer harm due to this gap. A second problem concerns blackletter antitrust law in dealing with price coordination through communication or facilitating practices; current doctrine requires that there be an anticompetitive “agreement” for there to be a violation of the Sherman Act for price fixing. Under standard models, even where oligopolists have independent incentives to price supracompetitively, they can often do better via an agreement; moreover, in other cases, competing firms can only achieve supracompetitive pricing by explicit collusion. In these cases, usually analyzed as a prisoner’s dilemma in which the Nash equilibrium is to “cheat” on the cartel, an agreement is required to avoid the inferior (from the price-fixers’ perspective) outcome. In order to find such an “agreement,” courts, government enforcers, and practitioners tend to focus on finding “intent,” efforts to sowing fear and distrust, and discovering a “meeting of the minds.” These standard inquiries derive from a more than a century-old embedded assumption that antitrust regulates sales by human actors; they will be a poor fit in addressing robo-sellers, which will function differently and which will likely not create the same kinds of evidence that these inquiries rely on. What can be done about the anticompetitive effects of robo-selling? This Article assesses several possible solutions, but find that they will be quite difficult to reconcile with current antitrust law. It conclude that, at least as a feasible second-best result, incorporating an evolving approach to robo-sellers may be a worthwhile expansion of the FTC’s ongoing regulatory program that has already begun target the competition and consumer protection aspects of consumer data collection by sellers. For example, the FTC has already begun to consider the effects of mass data collection and algorithmic processing on consumers from the perspective of disclosure and discrimination (both price and social); efficiencies should exist in broadening the inquiry to include effects on price coordination and cartel behavior.	2015	Salil K Mehra	Antitrust and the Robo-Seller: Competition in the Time of Algorithms
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms as Illegal Agreements	Despite the increased transparency, connectivity, and search abilities that characterize the digital marketplace, the digital revolution has not always yielded the bargain prices that many consumers expected. What is going on? Some researchers suggest that one factor may be coordination between the algorithms used by suppliers to determine trade terms. Simple coordination-facilitating algorithms are already available off the shelf, and such coordination is only likely to become more commonplace in the near future. This is not surprising. If algorithms offer a legal way to overcome obstacles to profit-boosting coordination, and create a jointly profitable status quo in the market, why should suppliers not use them? In light of these developments, seeking solutions – both regulatory and market-driven – is timely and essential. While current research has largely focused on the concerns raised by algorithmic-facilitated coordination, this article takes the next step, asking to what extent current laws can be fitted to effectively deal with this phenomenon. To meet this challenge, this article advances in three stages. The first part analyzes the effects of algorithms on the ability of competitors to coordinate their conduct. While this issue has been addressed by other researchers, this article seeks to contribute to the analysis by systematically charting the technological abilities of algorithms that may affect coordination in the digital ecosystem in which they operate. Special emphasis is placed on the fact that the algorithms is a “recipe for action”, which can be directly or indirectly observed by competitors. The second part explores the promises as well as the limits of market solutions. In particular, it considers the use of algorithms by consumers and off-the-grid transactions to counteract some of the effects of algorithmic-facilitated coordination by suppliers. The shortcomings of such market solutions lead to the third part, which focuses on the ability of existing legal tools to deal effectively with algorithmic-facilitated coordination, while not harming the efficiencies they bring about. The analysis explores three interconnected questions that stand at the basis of designing a welfare-enhancing policy: What exactly do we wish to prohibit, and can we spell this out clearly for market participants? What types of conduct are captured under the existing antitrust laws? And is there justification for widening the regulatory net beyond its current prohibitions in light of the changing nature of the marketplace? In particular, the article explores the application of the concepts of plus factors and facilitating practices to algorithms. The analysis refutes the Federal Trade Commission’s acting Chairwoman’s claim that current laws are sufficient to deal with algorithmic-facilitated coordination.	2018	Michal Gal	Algorithms as Illegal Agreements
AI facilitated collusion Principles for AI Regulation and Competition Law	Artificial Intelligence & Collusion: When Computers Inhibit Competition	The development of self-learning and independent computers has long captured our imagination. The HAL 9000 computer, in the 1968 film, 2001: A Space Odyssey, for example, assured, “I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do.” Machine learning raises many challenging legal and ethical questions as to the relationship between man and machine, humans’ control -- or lack of it -- over machines, and accountability for machine activities. While these issues have long captivated our interest, few would envision the day when these developments (and the legal and ethical challenges raised by them) would become an antitrust issue. Sophisticated computers are central to the competitiveness of present and future markets. With the accelerating development of AI, they are set to change the competitive landscape and the nature of competitive restraints. As pricing mechanisms shift to computer pricing algorithms, so too will the types of collusion. We are shifting from the world where executives expressly collude in smoke-filled hotel rooms to a world where pricing algorithms continually monitor and adjust to each other’s prices and market data. Our paper addresses these developments and considers the application of competition law to an advanced ‘computerised trade environment.’ After discussing the way in which computerised technology is changing the competitive landscape, we explore four scenarios where AI can foster anticompetitive collusion and the legal and ethical challenges each scenario raises.	2017	Ariel Ezrachi, Maurice E Stucke	Artificial Intelligence & Collusion: When Computers Inhibit Competition
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithms, Machine Learning, and Collusion	This paper discusses the question whether self-learning price-setting algorithms are able to coordinate their pricing behaviour to achieve a collusive outcome that maximizes the joint profits of the firms using these algorithms. While the legal literature generally assumes that algorithmic collusion is indeed possible and in fact very easy, the computer science literature on cooperation between algorithms as well as the economics literature on collusion in experimental oligopolies indicate that a coordinated and in particular tacitly collusive behaviour is in general rather difficult to achieve. Many studies have shown that some form of communication is of vital importance for collusion if there are more than two firms in a market. Communication between algorithms is also a topic in artificial intelligence research and some recent contributions indicate that algorithms may learn to communicate, albeit in a rather limited way. This leads to the conclusion that algorithmic collusion is currently much more difficult to achieve than often assumed in the legal literature and is therefore currently not a particularly important competitive concern. In addition, there are also several legal problems associated with algorithmic collusion, for example, questions of liability, of auditing and monitoring algorithms as well as enforcement. The limited resources of competition authorities should rather be devoted to more pressing problems as, for example, the abuse of dominant positions by large online-platforms.	2018	Ulrich Schwalbe	Algorithms, Machine Learning, and Collusion
AI facilitated collusion AI and market power	Autonomous Algorithmic Collusion: Q-Learning Under Sequential Pricing	Prices are increasingly set by algorithms. One concern is that intelligent algorithms may learn to collude on higher prices even in the absence of the kind of coordination necessary to establish an antitrust infringement. However, exactly how this may happen is an open question. I show how in simulated sequential competition, competing reinforcement learning algorithms can indeed learn to converge to collusive equilibria when the set of discrete prices is limited. When this set increases, the algorithm considered increasingly converges to supra-competitive asymmetric cycles. I show that results are robust to various extensions and discuss practical limitations and policy implications.	2021	Timo Klein	Autonomous Algorithmic Collusion: Q-Learning Under Sequential Pricing
AI facilitated collusion Principles for AI Regulation and Competition Law	When Machines Learn to Collude: Lessons from a Recent Research Study on Artificial Intelligence	From Professors Maurice Stucke and Ariel Ezrachi’s Virtual Competition published a year ago, to speeches by the Federal Trade Commission Commissioner Terrell McSweeny and Acting Chair Maureen K. Ohlhausen, to an entire issue of a recent CPI Antitrust Chronicles, and a conference hosted by Organisation for Economic Co-operation and Development (OECD) in June this year, there has been an active and ongoing discussion in the antitrust community about computer algorithms. In this note, I briefly summarize the current views and concerns in the antitrust and artificial intelligence (AAI) literature pertaining to algorithmic collusion and then discuss the insights and lessons we could learn from a recent AI research study. As I argue in the article, not all assumptions in the antitrust scholarship have empirical support at this point.	2020	Ai Deng	When Machines Learn to Collude: Lessons from a Recent Research Study on Artificial Intelligence
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithmic Pricing: What Implications for Competition Policy?	Pricing decisions are increasingly in the “hands” of artificial algorithms. Scholars and competition authorities have voiced concerns that those algorithms are capable of sustaining collusive outcomes more effectively than human decision makers. If this is so, then our traditional policy tools for fighting collusion may have to be reconsidered. We discuss these issues by critically surveying the relevant law, economics and computer science literatures.	2018	Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò, Sergio Pastorello	Algorithmic Pricing: What Implications for Competition Policy?
AI facilitated collusion Principles for AI Regulation and Competition Law	Developing Competition Law for Collusion by Autonomous Price-Setting Agents	After arguing that collusion by software programs which choose pricing rules without any human intervention is not in violation of section 1 of the Sherman Act, the paper offers a path towards making collusion by autonomous agents unlawful.	2017	Joseph E Harrington Jr	Developing Competition Law for Collusion by Autonomous Price-Setting Agents
AI facilitated collusion AI and market power	Algorithmic Collusion in Electronic Markets: The Impact of Tick Size	We characterise the stochastic interaction of learning algorithms as a deterministic system of differential equations to understand their long-term behaviour in a repeated game. In a symmetric bimatrix repeated game, we prove that the dynamics of many learning algorithms converge to the outcomes of pure strategy Nash equilibria of the stage game. In market making, we show that the algorithms tacitly collude to extract rents and tick size (coarseness of price grid) matters: a large tick size obstructs competition, while a smaller tick size lowers trading costs for liquidity takers, but slows the speed of convergence to an equilibrium.	2022	Álvaro Cartea, Patrick Chang, José Penalva	Algorithmic Collusion in Electronic Markets: The Impact of Tick Size
AI facilitated collusion AI and market power	Algorithms, Artificial Intelligence and Simple Rule Based Pricing	The increasingly popular automated pricing strategies in e-commerce can be broadly categorized into two forms: simple rule-based algorithms, such as undercutting the lowest price, and more sophisticated artificial intelligence (AI) powered algorithms, like reinforcement learning (RL). RL algorithms are particularly appealing for pricing due to their potential ability to autonomously learn an optimal policy and adapt to changes in competitors' strategies and market conditions. Despite the common belief that RL algorithms hold a significant advantage over rule-based strategies, our extensive experiments, conducted under both a canonical Logit demand environment and a more realistic non-sequential search structural demand model, demonstrate that when competing against RL pricing algorithms, simple rule-based algorithms can lead to higher prices and benefit all sellers, compared to scenarios where multiple RL algorithms compete against each other. Theoretical analysis in a simplified setting yields consistent results. Our research sheds new light on the effectiveness of automated pricing algorithms and their interactions in competitive markets, providing practical insights for retailers in selecting appropriate pricing strategies.	2025	Qiaochu Wang, Yan Huang, Param Vir Singh, Kannan Srinivasan	Algorithms, Artificial Intelligence and Simple Rule Based Pricing
AI facilitated collusion	Platform Design When Sellers Use Pricing Algorithms	We investigate the ability of a platform to design its marketplace to promote competition, improve consumer surplus, and increase its own payoff. We consider demand-steering rules that reward firms that cut prices with additional exposure to consumers. We examine the impact of these rules both in theory and by using simulations with artificial intelligence pricing algorithms (specifically Q-learning algorithms, which are commonly used in computer science). Our theoretical results indicate that these policies (which require little information to implement) can have strongly beneficial effects, even when sellers are infinitely patient and seek to collude. Similarly, our simulations suggest that platform design can benefit consumers and the platform, but that achieving these gains may require policies that condition on past behavior and treat sellers in a non-neutral fashion. These more sophisticated policies disrupt the ability of algorithms to rotate demand and split industry profits, leading to low prices.	2022	Justin Johnson, Andrew Rhodes, Matthijs R Wildenbeest	Platform Design When Sellers Use Pricing Algorithms
AI facilitated collusion	Algorithmic Pricing and Competition: Empirical Evidence from the German Retail Gasoline Market	Economic theory provides ambiguous and conflicting predictions about the association between algorithmic pricing and competition. In this paper we provide the first empirical analysis of this relationship. We study Germany’s retail gasoline market where algorithmic-pricing software became widely available by mid-2017, and for which we have access to comprehensive, highfrequency price data. Because adoption dates are unknown, we identify gas stations that adopt algorithmic-pricing software by testing for structural breaks in markers associated with algorithmic pricing. We find a large number of station-level structural breaks around the suspected time of large-scale adoption. Using this information we investigate the impact of adoption on outcomes linked to competition. Because station-level adoption is endogenous, we use brand headquarter-level adoption decisions as instruments. Our IV results show that adoption increases margins by 9%, but only in non-monopoly markets. Restricting attention to duopoly markets, we find that market-level margins do not change when only one of the two stations adopts, but increase by 28% in markets where both do. These results suggest that AI adoption has a significant effect on competition.	2020	Stephanie Assad, Robert Clark, Daniel Ershov, Lei Xu	Algorithmic Pricing and Competition: Empirical Evidence from the German Retail Gasoline Market
AI facilitated collusion	Algorithmic Pricing in Multifamily Rentals: Efficiency Gains or Price Coordination?	This paper empirically evaluates the impact of algorithmic pricing on the U.S. multifamily rental market. We hand-collect data on management company adoption decisions of algorithmic pricing and combine it with a comprehensive database of building-level rents and occupancy from 2005 to 2019. We find strong evidence that algorithmic pricing helps building managers set prices that are more responsive to market conditions, with adopters lowering rents more rapidly than non-adopters during economic downturns. We also find that average rents are higher and average occupancies are lower in markets with greater algorithmic penetration during periods of economic recovery. Then, we estimate a structural model of housing demand to test for "algorithmic coordination." Compared to a model of own profit maximization, our pair-wise tests favor a model of joint profit maximization among adopters of the same software. We estimate that the coordination channel results in an average markup increase of $25 per unit per month, impacting about 4.2 million units nationwide. Our findings have important implications for regulators and policymakers concerned about the potential risks and trade-offs of algorithmic pricing.	2024	Sophie Calder-Wang, Gi Heung Kim	Algorithmic Pricing in Multifamily Rentals: Efficiency Gains or Price Coordination?
AI facilitated collusion	Sustainable and Unchallenged Algorithmic Tacit Collusion	Algorithmic collusion is a hot topic within antitrust circles in Europe, US and beyond. But some economists downplay algorithmic collusion as unlikely, if not impossible. This paper responds to these criticisms by pointing to new emerging evidence and the gap between law and this particular economic theory. We explain why algorithmic tacit collusion isn’t only possible, but warrants the increasing concerns of many enforcers.	2020	Ariel Ezrachi, Maurice E Stucke	Sustainable and Unchallenged Algorithmic Tacit Collusion
AI facilitated collusion	Artificial Intelligence, Algorithmic Pricing and Collusion	Pricing algorithms are increasingly replacing human decision making in real marketplaces. To inform the competition policy debate on possible consequences, we run experiments with pricing algorithms powered by Artificial Intelligence in controlled environments (computer simulations). In particular, we study the interaction among a number of Q-learning algorithms in the context of a workhorse oligopoly model of price competition with Logit demand and constant marginal costs. We show that the algorithms consistently learn to charge supra-competitive prices, without communicating with each other. The high prices are sustained by classical collusive strategies with a finite punishment phase followed by a gradual return to cooperation. This finding is robust to asymmetries in cost or demand and to changes in the number of players.	2019	Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò, Sergio Pastorello	Artificial Intelligence, Algorithmic Pricing and Collusion
AI facilitated collusion	Algorithmic-facilitated Coordination	The use of algorithms in digital markets brings about many benefits. They offer consumers the ability to compare online offers in a more efficient and sophisticated manner, thereby enabling consumers to enjoy lower-priced products, or products that better fit their preferences.1 They enable suppliers to more quickly and efficiently analyze large amounts of data which is updated in real time on market conditions, thereby allowing suppliers to better and more quickly respond to consumer demand, to better allocate production and marketing resources, and to save on human capital. To do so, algorithms perform a myriad of tasks including sorting through data, organizing it, and making decisions based on the data collected with regard to multiple issues.. These advantages, researchers claim, are currently threatened by algorithmicfacilitated coordination. As the argument goes, algorithms make coordination among suppliers - both implicit or tacit- much easier and quicker than ever before. Coordination can be sustained at lower levels of concentration; and firms can more quickly and easily detect and punish deviations from the coordinated equilibrium, thereby reducing incentives for shirking. As our assumptions about which market conditions must exist for firms to coordinate are altered, the number of red flags that are raised rises. Ezrachi and Stucke suggest in their seminal work on virtual competition that this effect is so strong that is the end of competition as we know it. Should, indeed, algorithms facilitate tacit coordination in markets not otherwise prone to it, we need to explore which tools – either market-based or regulatory – can be used, if at all, in order to reduce the negative welfare effects of algorithmic coordination among competitors. If some of the assumptions that stand at the basis of the current rule under which tacit collusion is not considered an “agreement in restraint of trade” do not hold anymore, such as the assumption that collusion can generally only be reached in highly concentrated markets, it is time to determine whether our laws are fit to deal with the digitized world; whether we are looking under the lamp while most of the occurrence in the real world is happening outside its scope of light. To further use this metaphor in the remedial stage- can we widen the scope of our existing laws by simply using a stronger light bulb in the same lamp, or do we need to create a new source of light altogether? Accordingly, this paper focuses on three issues that arise from this technological challenge. First, it explores the effects of algorithms on the ability of suppliers to coordinate their conduct. Second, it briefly explores the promises as well as the limits of market solutions to welfare-reducing algorithmic coordination. The third part explores the ability of existing legal and regulatory tools to deal effectively with algorithmic facilitated coordination. Such tools can be complementary or provide at least some viable alternative for the possible failure of market-based solutions to deal with algorithmic facilitated coordination. The analysis explores three interconnected questions that stand at the basis of designing a welfare-enhancing policy towards the use of coordination facilitating algorithms: Do algorithms that facilitate coordination fulfil the requirement of “an agreement”, and if so- under which conditions?; what exactly do we wish to prohibit and can we spell it out clearly for market participants?; and is there justification for widening the regulatory net beyond its current prohibitions, in light of the changing nature of the marketplace.	2017	OECD Michal Gal	Algorithmic-facilitated coordination
AI facilitated collusion	Algorithmic Collusion: Problems and Counter-Measures	In this paper, we explore how technological advancements have changed, and will continue to change, the dynamics of competition and subsequently the distribution of wealth in society. How algorithms may be used in stealth mode to stabilize and dampen market competition while retaining the façade of competitive environment. We first raised algorithmic tacit collusion in 2015. In 2016 we provided further context and analysis in our book, Virtual Competition: The Promise and Perils of the Algorithm Driven Economy. We illustrated how online tacit collusion may emerge when products are generally homogeneous and sellers do not benefit from brand recognition or loyalty, and when markets are transparent and concentrated	2017	OECD Ariel Ezrachi Maurice E Stucke	Algorithmic Collusion: Problems and Counter-Measures
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithmic Pricing and Market Coordination – Toward a Notion of ‘Collusive Risk’	Over the past couple of years, many competition and antitrust scholars have feared the dawn of ‘algorithmic collusion’. Some have thus suggested expanding the notions of ‘collusion’ and ‘agreement’ in order to capture such coordination. Rather than using an expansive reading of ‘collusion’, the author of this article suggests an approach that works with the core and original intent of Article 101(1) TFEU: the fostering of independent conduct and prevention of market coordination. It finds this to be doctrinally undisputed and also consistent with longstanding competition policy debates, as well as an egalitarian notion of price that lays the foundation of the free market economy. On this basis, and considering given uncertainties, an operational notion of ‘collusive risk’ is put forward	2020	Juliane Mendelsohn	Algorithmic Pricing and Market Coordination – Toward a Notion of ‘Collusive Risk’
AI facilitated collusion Principles for AI Regulation and Competition Law	Autonomous Algorithmic Price Coordination	Ordering either a cab, buying some food, or purchasing any article over the internet usually offers a price that you believe is most reasonable and (usually) cheaper than those offering similar services offline. What makes this different is the pricing strategy which decides the price of similar goods available online vs. offline. While fixing the correct price has often been a problem of economic theory, traditionally, it is decided according to the cost of the product and demand-supply of the commodity. However, the introduction of algorithms into this stream certainly helps firms set a perfect price, albeit with the help of a series of data and analytics. It is often said that computational tools help set a correct price and elevate consumer welfare, but a question or (better say) suspicion arises on the possibility of algorithms colluding and coordinating to decide the prices of your cab ride. This paper explores on such tendency of algorithms and checks the efficiency of underlying Indian legal framework and whether it is capable to handle such coordination.	2021	Prakhar Harit	Autonomous Algorithmic Price Coordination
AI facilitated collusion Principles for AI Regulation and Competition Law	Antitrust Law and Coordination Through Al-Based Pricing Technologies	Price is the core element of commercial transactions and an important parameter of competition. One of antitrust law’s aims is to ensure that market prices form under the laws of supply and demand, and not after the whims of monopolists or cartelists. Innovations in computer and data science have brought about pricing technologies that rely on advanced analytics or machine learning (ML) techniques, which could strengthen existing bargaining power disparities in part by supporting price coordination among competitors. Existing research establishes a theoretical framework for competitive harm through coordination, showing that pricing technologies can lead to near-cartel price levels while avoiding anti-cartel prohibitions. This contribution builds on that framework, taking into account up to date empirical, game-theoretic, and computer science literature on pricing technologies to produce a taxonomy of those technologies. We then employ a comparative approach to identify the legal effects of various pricing technologies at a more granular level under EU and US antitrust law. The contribution supports greater understanding between economists and policy- makers regarding the analysis and treatment of AI-based pricing technologies.	2024	Maria José Schmidt-Kessen, Max Huffman	Antitrust Law and Coordination Through Al-Based Pricing Technologies
AI facilitated collusion Principles for AI Regulation and Competition Law	Algorithmic Collusion & Its Implications for Competition Law and Policy	Algorithms are increasingly employed by businesses as an integral part of their business models given the availability of big data and breakthroughs in artificial intelligence technology and application. While the competition landscape has shifted to a digital environment, questions have been raised as to whether the traditional competition policy tools, formulated in the analogue era, may nevertheless remain relevant in addressing algorithmic anti-competitive practices. In particular, the unilateral use of algorithms and algorithmic tacit collusion raises enforcement challenges for the competition authorities because of the inability of existing ex-post measures to adequately address these algorithmic anti-competitive conducts. This paper discusses these issues through a critical analysis of the relevant law, economics and computer science literatures.	2019	Kenji Lee	Algorithmic Collusion & Its Implications for Competition Law and Policy
AI facilitated collusion Principles for AI Regulation and Competition Law	Collusion by Algorithm: Does Better Demand Prediction Facilitate Coordination Between Sellers?	We build a game-theoretic model to examine how better demand forecasting due to algorithms, machine learning and artificial intelligence affects the sustainability of collusion in an industry. We find that while better forecasting allows colluding firms to better tailor prices to demand conditions, it also increases each firm's temptation to deviate to a lower price in time periods of high predicted demand. Overall, our research suggests that, despite concerns expressed by policymakers, better forecasting and algorithms can lead to lower prices and higher consumer surplus.	2018	Jeanine Miklós-Thal, Catherine E Tucker	Collusion by Algorithm: Does Better Demand Prediction Facilitate Coordination Between Sellers?
AI facilitated collusion	Understanding AI Collusion and Compliance	Antitrust compliance scholarship, particularly with a focus on collusion, has been an area of study for some time. Changes in technology and the rise of artificial intelligence (AI) and machine-learning create new possibilities both for anti-competitive behavior and to aid in detection of such algorithmic collusion. To some extent, AI collusion takes traditional ideas of collusion and simply provides a technological overlay to them. However, in some instances, the mechanisms of both collusion and detection can be transformed using AI. This handbook chapter discusses existing theoretical and empirical work, and identifies research gaps as well as avenues for new scholarship on how firms or competition authorities might invest in AI compliance to improve detection of wrong doing. We suggest where AI collusion is possible and offer new twists to where prior work has not identified possible collusion. Specifically, we identify the importance of AI to address the “trust” issue in collusion. We also identify that AI collusion is possible across non-price dimensions, such as manipulated product reviews and ratings, and discuss potential screens involving co-movements of prices and ratings. We further emphasize that AI may encourage entry, which may limit collusive prospects. Finally, we discuss how AI can be used to help with compliance both at the firm level and by competition authorities.	2020	Justin Johnson Cornell, D Daniel Sokol	Understanding AI Collusion and Compliance
AI facilitated collusion	How Concerned Should We Be About Algorithmic Tacit Collusion? Comments on Calvano et al.	A recent study by four economists (Calvano et al, 2019) has undoubtedly refueled the public and academic interest in algorithmic tacit collusion. In this study, the researchers find that their “algorithms consistently learn to charge supra-competitive prices, without communicating with one another… this finding is robust to asymmetries in cost or demand, changes in the number of players, and various forms of uncertainty.” Not surprisingly, this research has since been reported in a number of outlets including the Wall Street Journal, Financial Times, MIT Tech Review, among others. What does the study say about the likelihood of algorithmic collusion in the real world? Does it fundamentally change how we should think about tacit collusion in general? These are the questions I address in this note.	2019	Ai Deng	How Concerned Should We Be About Algorithmic Tacit Collusion? Comments on Calvano et al.
AI facilitated collusion AI exclusion and exploitation	Price Discrimination-Driven Algorithmic Collusion: Platforms for Durable Cartels	Algorithmic competition has arrived. With it has come the specter of algorithmic collusion – rapid detection of co-conspirators’ defection via technologically enhanced price monitoring and setting capability can encourage anticompetitive collusion. Strikingly, the ability to track consumers’ willingness-to-pay and price discriminate among them may synergize with algorithmic collusion into something antitrust scholars had previously thought impossible: stable cartels. In particular, consumer-facing digital platforms increasingly can determine consumers’ individual willingness to pay. Doing this allows them to deploy sophisticated forms of price discrimination, and thereby effect large welfare transfers from consumers to producers. This Article is the first to describe and analyze the potential interaction between price discrimination and algorithmic collusion. Algorithm-driven platforms now knit together large numbers of previously-independent firms and agents; some platforms set the price these participating firms and agents will charge. Crucially, if the gains to producers from collusive price discrimination are big enough, a qualitative change may take place: participants may find that they are no longer are in a Prisoner’s Dilemma tempting them to undercut each other on price, but rather in a coordination game with a single, rational choice: keep their collusion going. This Article sets forth how this dynamic can produce agreements by competitors, facilitated by price-discriminating, price-setting platforms that transfer wealth from consumers to producers – arguably a violation of Section 1 of the Sherman Act. Indeed, in contrast to the traditional view that firms need to first obtain Section 2 monopoly power, and only then can implement price discrimination, the model presented here shows the causation can run the other way: The ability to price discriminate effectively can drive the joint maintenance of monopoly power by colluding competitors. This dynamic takes on new urgency as more and more commerce shifts to the Internet and smartphone apps, a trend that has been accelerated by the COVID-19 pandemic and its associated acceleration of the shift to e-commerce. Potential solutions to this problem will be complicated by antitrust law’s current relegation of price discrimination to the dead letter office – no Federal Trade Commission complaint under the Robinson-Patman Act, the main relevant statute, has been brought this century. Indeed, during the past decade, the most recent edition of the leading antitrust casebook in the U.S. deleted its section on price discrimination and the Act. This Article proposes three actions: (i) revive some enforcement against price discrimination, (ii) prioritize action against price discriminating platforms that inhibit switching by participants, including scrutinizing mergers between firms whose Big Data-based ability to gauge willingness-to-pay may, if combined, have negative ramifications for consumers, and (iii) factor price discrimination-driven algorithmic collusion into the current reevaluation of vertical restraints.	2021	Salil K Mehra	Price Discrimination-Driven Algorithmic Collusion: Platforms for Durable Cartels
AI facilitated collusion	Algorithmic Collusion and Algorithmic Compliance: Risks and Opportunities	Algorithms are becoming ubiquitous in our society. They are powerful and, in some cases, indispensable tools in today’s economy. In terms of the technology, we do not yet have AI sophisticated enough to, with a reasonable degree of certainty, reach autonomous tacit collusion in most real markets. This does not mean that we should ignore the potential risks. In fact, in their effort to design AIs that can learn to cooperate with each other and with humans for social good, AI researchers have shown that autonomous algorithmic coordination is possible. But there are also several positive takeaways from this research. For example, given the technical challenges, I argue that just like emails leave a trail of evidence when a cartel uses them to coordinate, a similar trail of evidence is likely present when collusive algorithms are being designed. The literature also gives us a good deal of insights about the types of design features and capabilities that could lead to algorithmic collusion. I highlight the role of algorithmic communication as a leading example and argue that these known collusive features should raise red flags even if collusion is ultimately reached autonomously by algorithms.	2021	Ai Deng	Algorithmic Collusion and Algorithmic Compliance: Risks and Opportunities
AI facilitated collusion	Algorithmic Collusion: Fear of the Unknown or too Smart to Catch?	Algorithms, which businesses use more and more to set pricing strategies with each passing day, could be pro-competitive and provide significant efficiencies. Depending on how firms use them, however, they can also potentially restrict competition and harm consumers. Scholars and enforcers debate on the right scope of the theory of harm as to using algorithms in critical competition parameters, particularly in pricing strategies. Yet, one question remains open: if firms use algorithms that may tacitly collude through machine learning - particularly deep learning technologies, will that alone be sufficient to hold them liable for a competition law infringement? This article discusses the current limits of the collective knowledge on this subject and explores what guidance could still be provided to businesses. The article argues that the solution does not lie in taking premature regulatory actions without sufficient empirical evidence to justify any shift away from the traditional concepts of tacit collusion, or without proper guidance for companies to avoid the risk of legal exposure. The need is evident for further research on how self-learning algorithms operate in real-life settings, which could start defining a “red zone” for businesses to watch out and for enforcers to focus their energy and resources.	2021	Gonenc Gurkaynak, Burcu Can, Sinem Uğur	Algorithmic Collusion: Fear of the Unknown or too Smart to Catch?
AI facilitated collusion	Artificial Intelligence and Anticompetitive Collusion: From the ‘Meeting of Minds’ Towards the ‘Meeting of Algorithms’?	Information technologies has affected so many aspects of daily life that algorithmic society is not considered science fiction anymore. When it comes to marketplaces and business strategies, it has been observed that a growing number of firms are using algorithms for dynamic price setting, thereby automatically adjusting their prices to changes in market conditions, including rivals’ prices. As a result, the diffusion of algorithmic pricing raises concerns for competition policy about the potential to enable collusion. Further, some policy makers and scholars are questioning the ability of existing antitrust tools to tackle effectively this new form of collusion. Indeed, current antitrust rules have been designed to deal with human facilitation of coordination requiring some form of mutual understanding among firms (‘meeting of the minds’). However, according to a strand of literature, algorithms could coordinate independently of human intervention and even autonomously learn to collude. Against this background, the paper aims at investigating whether current antitrust rules are suited to facing these new challenges, whether algorithmic interactions (‘meeting of algorithms’) could be treated similarly to a ‘meeting of minds’ or whether new regulatory tools are needed.	2021	Giuseppe Colangelo	Artificial Intelligence and Anticompetitive Collusion: From the ‘Meeting of Minds’ Towards the ‘Meeting of Algorithms’?
AI facilitated collusion	Artificial Collusion: Examining Supracompetitive Pricing by Q-Learning Algorithms	We examine concerns that pricing algorithms used by competitors would autonomously and systematically learn to collude at supra-competitive prices. Findings of high prices with Q-learning have recently raised that alarm. A detailed analysis of the inner workings of this algorithm type reveals, however, that it does not constitute autonomous algorithmic collusion and is unlikely to be a risk in practice. The `collusive equilibria' only exist by the construction of the state space, a substantial fraction of supra-competitive prices is not sustained by a reward-punishment scheme, and observing reward-punishment patterns need not imply a scheme. If there is convergence on collusive equilibria, it is intrinsically slow and any benefits are obtained on timescales irrelevant to the firms' stated objectives. Moreover, Q-learning algorithms are outperformed by the first alternative pricing algorithm. Our analysis gives criteria for practically relevant colluding pricing algorithms that would constitute a threat to competition. They likely require malign programming, intent and explicit coordination, that would show from the codes.	2025	Arnoud V den Boer, Janusz M Meylahn, Maarten Pieter Schinkel	Artificial Collusion: Examining Supracompetitive Pricing by Q-Learning Algorithms
AI facilitated collusion	What do we know about algorithmic collusion now	Algorithmic collusion has captured the attention of the global antitrust community for the past several years. Deng (2020) provided a comprehensive survey of the pertinent literature in economics and computer science and a critical discussion. Over the past three years, new insights have emerged from academic research. These new insights have not only deepened our understanding of the intricate relationship between algorithms and competition but also begun challenging some previous findings once considered compelling evidence supporting the plausibility of autonomous algorithmic tacit collusion. In this article, I discuss these new insights, with a focus on the following questions: - Can AIs have a "meeting of minds"? - Is price increase post AI adoption a clear indication of collusion? - What is the latest research in AI and Operational Research tell us about autonomous AI collusion? Didn't Calvano et al (2020) tell us that it is totally possible? - What is an example of collusive AI algorithm (collusion by design)? How does it work? - Which is more dangerous, dumb algorithms or smart algorithms? - Does the use of 3rd party algorithm almost certainly lead to softening of competition and/or collusion? - What should practitioner watch out for when evaluating the effect of algorithmic pricing?	2024	Ai Deng	What Do We Know About Algorithmic Collusion Now? New Insights from the Latest Academic Research
AI facilitated collusion	The Role of Secondary Algorithmic Tacit Collusion in Achieving Market Alignment	The antitrust risks associated with the use of the same hub’s pricing algorithm by many sellers are now well-accepted. But what if many rivals use several different hubs for dynamic pricing? The common assumption is that in such instances, competition among the pricing hubs would support competition among the sellers. However, in this paper we argue differently and introduce the concept of secondary algorithmic tacit collusion, which leads to anticompetitive effects, independent of the conditions on the primary market. This phenomenon may lead to the evils of price-fixing but on far a wider scale. Contrary to traditional tacit collusion, this aggregated form of collusion, through the use of algorithmic hub-and-spoke structures, can occur in markets with many competitors and with seemingly competitive dynamics. We outline how the combination of hub-and-spoke frameworks on the primary market and conscious parallelism on the secondary market for algorithmic pricing services can lead to secondary tacit collusion. Addressing its anticompetitive effects requires competition agencies to consider the interaction between price setters in the secondary markets, while taking note of the hub-and-spoke structures on the primary market.	2024	Ariel Ezrachi, Maurice E Stucke	The Role of Secondary Algorithmic Tacit Collusion in Achieving Market Alignment
AI facilitated collusion	Regulating Algorithms in The European Data-Driven Economy: The Role of Competition Law and Civil Liability	While showing the potential to make the market more competitive and efficient, algorithms are acknowledged to pose a challenge to competition law enforcement. This is because algorithmic tacit collusion does not amount to an outright agreement but is something more than mere market parallelism, which is normal in competitive markets. This essay reviews the economic and legal scholarship, the national and supranational case law and the supranational policy debate on this issue to explore if and how competition law can play a role in clarifying such a grey area, without discouraging technological innovation and economic development. In this regard, this essay finds that, while the case law has already addressed algorithms implementing explicit anticompetitive agreements under Article 101 TFEU, scholars fail to agree on how to tackle algorithmic tacit collusion. This has come under the radar of ongoing policy initiatives, such as the European Commission’s New Competition Tool initiative. Waiting for innovative regulatory and competition law solutions to better tackle algorithmic collusion, this essay proposes to use, as an alternative to Article 101 TFEU, the notion of collective abuse of dominant position under Article 102 TFEU. Finally, this essay considers how civil liability and private enforcement may contribute to competition law enforcement against algorithmic collusion.	2020	Andrea Parziale	Regulating Algorithms in The European Data-Driven Economy: The Role of Competition Law and Civil Liability
AI facilitated collusion	Eturas: Of Concerted Practices, Tacit Approval, and the Presumption of Innocence	The Court of Justice (ECJ) finds that the existence of a concerted practice under Article 101 TFEU may be presumed in case of dispatching an email (of which the addressees are aware) concerning restrictions of trading on a common digital platform, unless the respective addressees rebut that presumption.	2016	Catalin S Rusu	Eturas: Of Concerted Practices, Tacit Approval, and the Presumption of Innocence
AI facilitated collusion	Can Computers Form Cartels? About the Need for European Institutions to Revise the Concertation Doctrine in the Information Age	Traditionally, the European Commission and the European courts have considered that a concertation arises as soon as information is exchanged among competitors. That approach creates difficulties on information-based markets where computers and more generally machines systematically organise such exchanges and may thus give rise to allegations of cartel infringement for their operators, despite the absence of any fraudulous intention whatsoever on the part of the latter. For the authors, such development emphasizes the need, for the European institutions, to revisit their doctrine, and their jurisprudence, on the formation of anticompetitive coordination.	2016	Andreas Heinemann , Aleksandra Gebicka	Can Computers Form Cartels? About the Need for European Institutions to Revise the Concertation Doctrine in the Information Age
AI facilitated collusion	Die Begehung eines Wettbewerbsdelikts durch Empfang eines Rundschreibens	Der EuGH hat in der bemerkenswerten Entscheidung vgl. Urteil Eturas u. a., C‑74/14, EU:C:2016:42, BeckEuRS 2014, 416705 das Tatbestandsmerkmal der „aufeinander abgestimmten Verhaltensweise“ i. S.d Art. 101 AEUV weiter präzisiert. Der Gerichtshof bleibt mit diesem Judikat seiner restriktiven Linie bei der Auslegung des Art. 101 AEUV treu. Nunmehr kann bereits der Empfang eines elektronischen Rundschreibens mit wettbewerbswidrigem Inhalt die kartellrechtliche Haftung des empfangenden Unternehmens auslösen, vorausgesetzt es distanziert sich nicht öffentlich von der intendierten Verhaltensweise oder zeigt diese bei den Verfolgungsbehörden nicht an. Damit kommt auch die passive Begehung eines Wettbewerbsdelikts in Betracht. Diese Entscheidung ist daher für Unternehmen von großer Bedeutung, ist doch die Vermeidung von wettbewerbswidrigen Verhaltensweisen ein wichtiger Bestandteil ihrer Compliance-Bemühungen.	2016	Alexander Eufinger	Die Begehung eines Wettbewerbsdelikts durch Empfang eines Rundschreibens
AI facilitated collusion	Online sales of posters and frames	The CMA investigated a cartel relating to sales of posters and frames by 2 competing online sellers on Amazon’s UK website.	2016	Competitions and Markets Authority),	Online sales of posters and frames

AI facilitated manipulations

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
AI facilitated manipulations	Fake Sales and Ranking Algorithms in Online Retail Marketplace with Sponsored Advertising	We study the optimal algorithm decisions of a platform on ranking products sold by sellers--who may use fake sales to boost the rankings of their products--and the impact on consumers and sellers. We design a model of an online retail marketplace with competing sellers. The platform decides whether to tolerate fake sales and whether to rank its organic results based on sellers' qualities or popularities. The sellers decide whether to buy fake sales and how much to bid for sponsored advertising on the platform. We show a platform may strategically tolerate or even encourage popularity-boosting fake sales by a seller when the seller's quality level is extreme relative to competitors. With a low-quality seller, allowing fake sales may benefit the platform through reducing differentiation between sellers and intensifying competition for the sponsored ads (i.e., "feeding the puppy dog"). With a high-quality seller, it benefits the platform by increasing differentiation between the two sellers and softening price competition, which improves the platform's commission revenue (i.e., "feeding the fat cat"). Furthermore, fake sales may benefit consumers by increasing price competition and may also benefit competing sellers by reducing the high-quality seller's dominance.	2022	Fei Long, Yunchuan Liu	Fake Sales and Ranking Algorithms in Online Retail Marketplace with Sponsored Advertising
AI facilitated manipulations	Algorithmic Collusion Without Threats	There has been substantial recent concern that pricing algorithms might learn to ``collude.'' Supra-competitive prices can emerge as a Nash equilibrium of repeated pricing games, in which sellers play strategies which threaten to punish their competitors who refuse to support high prices, and these strategies can be automatically learned. In fact, a standard economic intuition is that supra-competitive prices emerge from either the use of threats, or a failure of one party to optimize their payoff. Is this intuition correct? Would preventing threats in algorithmic decision-making prevent supra-competitive prices when sellers are optimizing for their own revenue? No. We show that supra-competitive prices can emerge even when both players are using algorithms which do not encode threats, and which optimize for their own revenue. We study sequential pricing games in which a first mover deploys an algorithm and then a second mover optimizes within the resulting environment. We show that if the first mover deploys any algorithm with a no-regret guarantee, and then the second mover even approximately optimizes within this now static environment, monopoly-like prices arise. The result holds for any no-regret learning algorithm deployed by the first mover and for any pricing policy of the second mover that obtains them profit at least as high as a random pricing would -- and hence the result applies even when the second mover is optimizing only within a space of non-responsive pricing distributions which are incapable of encoding threats. In fact, there exists a set of strategies, neither of which explicitly encode threats that form a Nash equilibrium of the simultaneous pricing game in algorithm space, and lead to near monopoly prices. This suggests that the definition of ``algorithmic collusion'' may need to be expanded, to include strategies without explicitly encoded threats.	2024	Eshwar Ram Arunachaleswaran, Natalie Collina, Sampath Kannan, Aaron Roth, Juba Ziani	Algorithmic Collusion Without Threats
AI facilitated manipulations	Digital Market Manipulation	In 1999, Jon Hanson and Douglas Kysar coined the term “market manipulation” to describe how companies exploit the cognitive limitations of consumers. For example, everything costs $9.99 because consumers see the price as closer to $9 than $10. Although widely cited by academics, the concept of market manipulation has had only a modest impact on consumer protection law. This Article demonstrates that the concept of market manipulation is descriptively and theoretically incomplete, and updates the framework of the theory to account for the realities of a marketplace that is mediated by technology. Today’s companies fastidiously study consumers and, increasingly, personalize every aspect of the consumer experience. Furthermore, rather than waiting for the consumer to approach the marketplace, companies can reach consumers anytime and anywhere. The result of these and related trends is that firms can not only take advantage of a general understanding of cognitive limitations, but can uncover, and even trigger, consumer frailty at an individual level. A new theory of digital market manipulation reveals the limits of consumer protection law and exposes concrete economic and privacy harms that regulators will be hard-pressed to ignore. This Article thus both meaningfully advances the behavioral law and economics literature and harnesses that literature to explore and address an impending sea change in the way firms use data to persuade.	2014	Ryan Calo	Digital Market Manipulation’
AI facilitated manipulations	Amazon's Pricing Paradox	Antitrust scholars have widely debated the paradox of Amazon seemingly wielding monopoly power while charging low prices to consumers. A single company’s behavior thereby helped spark a vibrant intellectual conversation as scholars debated why Amazon’s prices were so low, whether enforcers should intervene, and, eventually, how the field of antitrust should be reformed. One of the main sources of agreement in these and other scholarly conversations has long been that Amazon charges low prices. This Article challenges that assumption by demonstrating that Amazon customers may pay significantly higher prices than is commonly understood due to strategies that do not necessarily depend on monopoly power. More importantly, unraveling the disconnect between perception and reality yields broader insights. One of the reasons why perceptions of Amazon’s pricing have remained disconnected from reality is that conversations about regulating Amazon have paid inadequate attention to behavioral economics. Behavioral economics reveals how the company leverages its sophisticated algorithms, large datasets, and dark patterns to build a marketplace of consumer misperception by, for instance, making it difficult for consumers to find the low-priced items. Such practices undermine the goals of competition, in the economic sense of the word. But these practices have traditionally been the focus of consumer law rather than antitrust. Indeed, the longstanding inattention to these consumer law-related behavioral pricing practices raises the question of whether scholars have been incorrectly describing Amazon’s prices as low. Amazon may offer many products at low, competitive prices, but by exploiting consumers’ behavioral biases, Amazon may prevent a substantial number of consumers from finding those low prices. Thus, a behavioral consumer lens is necessary to see that what was originally framed as an antitrust paradox is better viewed as a more general pricing paradox. A company perceived as offering low prices may have been instead manipulating consumers to pay more. To see the full set of concrete legal solutions for promoting competition in Amazon’s marketplace and beyond, it is important to move consumer law out of antitrust’s shadow. Consumer law interventions include mandating information disclosures by Amazon to empower artificially intelligent digital intermediaries that could help lower consumers’ search costs. Lawsuits based in unfair or deceptive acts or practices are also possible. Consumer law and antitrust law operating at full force offer the best chance for ushering in an era of “open retail” in which digital markets remain competitive and adequately serve consumers.	2023	Rory Van Loo, Nikita Aggarwal	Amazon's Pricing Paradox
AI facilitated manipulations AI and market power	Repricing Algorithms in E-Commerce	In this paper we explore models of repricing automation -- that is, of price adjustments implemented by an algorithm in response to changes in demand, inventory, or competitors’ prices. Our setting is a typical online multi-seller platform (e.g., Amazon, eBay) where competing firms sell products differentiated by a single quality dimension (e.g., seller reputation) and where quality is proxied by an observable metric (e.g., rating score). We study the performance of different repricing algorithms as compared with equilibrium prices and also analyze the robustness of those algorithms. In particular, we investigate the possible “gaming” of the automated price response by a strategic seller with perfect knowledge of its competitor’s repricing scheme. Our analysis affirms the reasonableness of the simple structure exhibited by most repricing rules observed in practice yet also identifies their downsides.	2015	Dana Popescu	Repricing Algorithms in E-Commerce

AI mergers and cooperation

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
AI mergers and cooperation AI facilitated collusion	Mergers, Acquisitions and Algorithms in an Algorithmic Pricing World	This paper considers whether the increased use of pricing algorithms presents issues that existing merger control frameworks and practices are inadequate to address, particularly in relation to pre-merger disclosure of pricing algorithms and the suitability of current tests, frameworks and remedies for addressing coordinated effects. These questions are addressed through a comparative and critical analysis of the merger control regimes and practices of the European Union, United Kingdom, and Australia. It finds that whilst such regimes and practices are broadly adequate for dealing with algorithmic transactions, there are nonetheless potential areas for improvement. Disclosure of a pricing algorithm may contravene prohibitions on the sharing of competitively sensitive information. As such, merger parties may need to rethink aspects of their usual due diligence procedures. Pricing algorithms may also increase the potential for coordinated effects to arise in some markets that would ordinarily have been considered too complex, asymmetrical, opaque, or insufficiently concentrated for tacit collusion to occur, or be used to retaliate more effectively against deviations from a coordinated equilibrium, or to raise the height of barriers to entry. Competition authorities may therefore need to amend their standard approach to investigating and assessing coordinated effects, as well as their traditional approach to remedies.	2022	Michael Coutts	Mergers, Acquisitions and Algorithms in an Algorithmic Pricing World
AI mergers and cooperation AI and market power	When Antitrust Becomes Pro-Trust: The Digital Deformation of U.S. Competition Policy	There is a delicate balance between consolidation and competition in any industry. In theory, mergers and acquisitions can allow firms to achieve economies of scale and scope. However, when concentration reaches a certain level, two distinct anti-competitive effects can emerge. First, within an industry, firms may feel pressure to grow simply to keep up with rivals. (When, for instance, the top two firms in an industry merge, the third largest one may quickly search out possible acquisition targets to keep up.) Second, the largest firm or firms in a very concentrated sector may use their pricing power to earn profits that allow them to expand outside the sector and take over firms in adjacent sectors	US	Frank Pasquale	When Antitrust Becomes Pro-Trust: The Digital Deformation of U.S. Competition Policy’
AI mergers and cooperation	Complex Antitrust Harm in Platform Markets	Should the FTC have allowed Zillow to acquire its foremost rival, Trulia? It is increasingly well-accepted that digital platforms tend toward dominance in their immediately adjacent relevant-product markets. Google, for example, has long held a majority share of the markets for general-search results and advertising, prompting antitrust and competition-law scrutiny of its conduct. But some digital platforms also possess the ability and incentive to increase concentration in seemingly removed, though related, markets. Complex platforms can harness the power of reputational mechanisms to steer their users toward favored third-party suppliers. A search engine, for example, might (under certain conditions identified by this article) rationally steer its users toward particular sellers of real-world products like restaurant meals, home goods, etc. This type of steering forecloses competition in markets not immediately adjacent to the platform itself. Current antitrust analyses focus solely on harm in "relevant markets," overlooking potential harm in "related markets." The Zillow-Trulia merger illustrates how related-market harm might occur. Post-deal statements by company executives indicate that the FTC's clearance of the merger may have constituted a false negative, and that the merged firm may be increasing concentration in local real-estate agent markets. Related-market harm is inefficient and reduces consumer welfare. This article contends that future merger and conduct analyses should take seriously the possibility of such harm.	2017	John M Newman	Complex Antitrust Harm in Platform Markets
AI mergers and cooperation AI and market power	Why the Google Books Settlement is Procompetitive, Journal of Legal Analysis	Although the Google Books Settlement has been criticized as anticompetitive, I conclude that this critique is mistaken. For out-of-copyright books, the settlement procompetitively expands output by clarifying which books are in the public domain and making them digitally available for free. For claimed in-copyright books, the settlement procompetitively expands output by clarifying who holds their rights, making them digitally searchable, allowing individual digital display and sales at competitive prices each rightsholder can set, and creating a new subscription product that provides digital access to a near-universal library at free or competitive rates. For unclaimed in-copyright books, the settlement procompetitively expands output by helping to identify rightsholders and making their books saleable at competitive rates when they cannot be found. The settlement does not raise rival barriers to offering any of these books, but to the contrary lowers them. The output expansion is particularly dramatic for commercially unavailable books, which by definition would otherwise have no new output.	2010	Einer Elhauge	‘Why the Google Books Settlement is Procompetitive, Journal of Legal Analysis
AI mergers and cooperation	The Sharing Economy Meets the Sherman Act: Is Uber a Firm, a Cartel or Something in Between?	The sharing economy is a new industrial structure that is made possible by instantaneous internet communication and changes in life, work, and purchasing habits of individual entrepreneurs and consumers. Antitrust law is an economic regulatory scheme dating to 1890 (in the United States) and designed to address centrally controlled concentrations of economic power and threats that those concentrations would operate to contravene both consumer interest and economic efficiency. Antitrust needs reenvisioning and careful application to accommodate a modern enterprise structure in which thousands or millions of independent contractors joint forces to provide a service by agreement among themselves. The success of Uber, Airbnb, and other sharing economy firms, and the consumer benefits those firms promise, show both how difficult and how important that reenvisioning can be.	2017	Mark Anderson, Max Huffman	The Sharing Economy Meets the Sherman Act: Is Uber a Firm, a Cartel or Something in Between?
AI mergers and cooperation AI facilicated collusion AI and market power	Algorithms, AI and Mergers	Algorithms, especially those based on artificial intelligence, play an increasingly important role in our economy. They are used by market participants to make pricing, output, quality, and inventory decisions; to predict market entry, expansion, and exit; and to predict regulatory moves. In a growing number of jurisdictions, algorithms are also used by regulators to detect and analyze anti-competitive conduct. This game-changing switch to (semi-)automated decision-making has the potential to reshape market dynamics. While the effect of algorithms on coordination between competitors has been a focus of attention, and scholarly work on their effects on unilateral conduct is beginning to accumulate, merger control issues have been undertreated. Accordingly, this article focuses on such issues. The article identifies six main functions of algorithms that may affect market dynamics: collection and ordering of data; improving the ability to use existing data; reducing the need for data, for in-stance by generating synthetic data; monitoring; predicting, to deter-mine how different types of conduct, including mergers, are likely to affect market conditions; and decision-making. The article demonstrates how such algorithms can exacerbate anti-competitive conduct with respect to both unilateral and coordinated effects. Towards this end, seven scenarios are explored: collusion, oligopolistic coordination, high unilateral prices, price discrimination, predation, selective pricing (in which a buyer offers a higher price to some suppliers in an aggressive bid for an input), and reducing the interoperability of datasets. For each scenario, we analyze how the market conditions necessary for such conduct are affected by algorithms. These findings are then translated into merger policy. Algorithms are shown to affect substantive as well as institutional features of merger control. Algorithms also challenge some of the assumptions that are ingrained in merger control, suggesting that a more informed approach to some algorithmic-related mergers is appropriate.	2023	Michal Gal, Daniel L Rubinfeld	Algorithms, AI and Mergers
AI mergers and cooperation AI and market power	Through the Looking Glass: AI, Mergers and the Role of Competition Law in Digital Governance	In this paper, I consider the critical question of the role of competition policy in digital governance. To do so, I consider what many see as a microcosm of the promise and the risks associated with our digital world – the progressive and rapidly expanding place of artificial intelligence within our economy – against the backdrop of what is considered the paradigmatic example of market regulation and competition policy – merger review. I situate the analysis in Canadian law, where the adaptation of regulation to the digital economy is still in its infancy, with the goal of identifying promising policy proposals to guide the essential legislative and regulatory process that lies ahead. Using a tailored concept of AI that focuses on its constituent economic inputs, I examine more closely how economic incentives and business practices informed by AI interact with traditional models of assessing the competitive impacts of market power. This provides a sharper lens for considering whether and how much competition law ought to be adapted in response to the realities of the new economy because it shows the practical impossibility of neatly separating the economic impact of AI from its non-economic impacts. I contend that responding to this new reality of blurred lines will require a serious re-examination of the limited footprint of Canadian competition law beyond its current ambit. My assessment is that given Canada’s legal and political structure and existing enforcement capacity limits, competition law will have to assume a greater role in digital governance, at least in the short term. Looking further ahead, I argue that competition policy must be part of a whole-of-government governance structure where relevant policy areas are coordinated to ensure Canada is well-positioned to respond to both existing and future issues related to the digital transformation, particularly those that cross sectorial lines.	2022	Jennifer A Quaid	Through the Looking Glass: AI, Mergers and the Role of Competition Law in Digital Governance
AI mergers and cooperation Principles for AI Regulation and Competition Law	The Sharing Economy Meets the Sherman Act: Is Uber a Firm, a Cartel, or Something in Between?	The sharing economy is a new industrial structure that is made possible by instantaneous internet communication and changes in life, work, and purchasing habits of individual entrepreneurs and consumers. Antitrust law is an economic regulatory scheme dating to 1890 (in the United States) and designed to address centrally controlled concentrations of economic power and threats that those concentrations would operate to contravene both consumer interest and economic efficiency. Antitrust needs reenvisioning and careful application to accommodate a modern enterprise structure in which thousands or millions of independent contractors joint forces to provide a service by agreement among themselves. The success of Uber, Airbnb, and other sharing economy firms, and the consumer benefits those firms promise, show both how difficult and how important that reenvisioning can be.	2017	Mark Anderson, Max Huffman	The Sharing Economy Meets the Sherman Act: Is Uber a Firm, a Cartel, or Something in Between?
AI mergers and cooperation AI exclusion and exploitation	The Role of Data Protection and Privacy Law in Personally Identifiable Information – Driven Mergers from the EU Merger Perspective	This chapter aims to review data-driven mergers including, but not limited to, major conglomerates involving large scale of individual user data, known as ‘big data’, by Facebook (WhatsApp), Microsoft (Yahoo!, Skype and LinkedIn), Google (Double Click), TomTom (Tele Atlas), Publicis/Omnicon, Telefonica/Vodafone UK, and so on. These mergers have been unconditionally cleared based on the traditional law and economic analysis of mergers, known as a ‘significant impediment to effective competition’ legal test. The test disregards public policy concerns, including the economics of privacy, i.e., data analytics; data sharing with third parties, e.g., publishers or retailers; and data selling. The chapter draws on previous research on the rise of big data and the loss of privacy, which sheds light inter alia on the ineffectiveness of the data, consumer and competition rules and on the intrusive privacy policies of the various digital platforms. This chapter argues that the current assessment of mergers has to activate the public policy clause and to consider the economic implications of privacy following a merger. No merger should be unconditionally cleared if it involves a large amount of users’ data. The chapter arrives at the conclusion that the new data protection framework is insufficiently robust. The contract theory of informed consent associated with the potential of sharing anonymised and/or aggregated data means that digital platforms are able to exploit data protection loopholes and abuse users’ trust in digital platforms. In addition, the chapter looks at the treatment of innovative digital platforms from the perspective of Schumpeterian economics and therefore identifies the fallacy of too great a reliance on ephemeral market shares. It discusses more critically the expectation of a robust and coherent theory of harm to consumers in the context of digital markets.	2020	Anca D Chirita	The Role of Data Protection and Privacy Law in Personally Identifiable Information – Driven Mergers from the EU Merger Perspective
AI mergers and cooperation AI facilitated collusion	Revising the U.S. DOJ-FTC Horizontal Merger Guidelines – Accounting for Algorithmic Coordination	This Comment is written in response to DOJ-FTC Request for Information on Merger Enforcement. We explain that the use of pricing algorithms based on artificial intelligence methodologies (hereinafter: "pricing algorithms"), by one or both parties, should be taken into account in the merger analysis. This is due to the fact that the use of such algorithms might substantially increase the possibility of explicit or tacit collusive behavior. We then suggest several ways in which merger review and the Horizontal Merger Guidelines can incorporate such effects.	2022	Michal Gal, Daniel L Rubinfeld	Revising the U.S. DOJ-FTC Horizontal Merger Guidelines – Accounting for Algorithmic Coordination

Principles for AI Regulation and Competition Law

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
Principles for AI Regulation and Competition Law	A General Framework for Analyzing the Effects of Algorithms on Optimal Competition Laws	Competition laws are influenced by economic presumptions regarding how markets operate. Such presumptions generally relate to how humans interact, such as how human decision-makerswhether acting as individuals or as a firm's agentsgather information, send signals, and deal with complex, uncertain, or fast-changing market environments. The exponential growth in the use of algorithms by market participants to perform a myriad of tasks is challenging such presumptions. The lowering of access barriers to real-time data on market conditions, coupled with semi-automated decision-making by sophisticated and autonomous robo-economicus, requires us to rethink the economic presumptions embedded in our laws. Indeed, as we show, in many cases, the application of existing legal presumptions to markets in which decisions are made by sophisticated algorithms operating on big data, increases the instances and the harms of false negatives and, although less frequently, false positives. While research thus far has focused on the effects of algorithms on specific types of competition rules, this article suggests a general framework for identifying such effects. We employ Decision Theory to help determine how competition laws should be optimally framed in the age of algorithmic decision-making. As we show, once the use of sophisticated AI-empowered algorithms is assumed, legal presumptions with regard to some types of conduct must be changed. We suggest a typology of six different effects, ranging from no effect at all to a need for new prohibitions. Our theoretical analysis is aided by real-world examples, including cases where the introduction of sophisticated algorithms affects the choice between rules versus standards, the content of the prohibition, or procedural rules. We hope our meta-analysis brings more clarity to a much-needed reboot of our regulatory framework in the age of algorithms	2024	Michal S Gal, Jorge Padilla	A General Framework for Analyzing the Effects of Algorithms on Optimal Competition Laws
Principles for AI Regulation and Competition Law AI mergers and cooperation	Reimagining Competition Policy in the Age of Sustainability and AI	• EU competition policy must be better aligned with green industrial objectives through flexible State aid and a stronger focus on innovation. • Sustainability is increasingly taken into account by competition authorities, but enforcement remains fragmented and uncertain. • The rise of AI and digital ecosystems challenges existing tools, calling for regulatory coherence and timely intervention. • Existing merger control rules are insufficient to address killer acquisitions, highlighting the need for new legal mechanisms and discretionary powers.	2025	Walid Chaiehloudj	Reimagining Competition Policy in the Age of Sustainability and AI
Principles for AI Regulation and Competition Law AI and Market Power AI exclusion and exploitation AI mergers and cooperation	KFTC's Generative AI and Competition Report: A Quick Overview	The Korea Fair Trade Commission (KFTC) recently released its first policy report on generative AI, titled “Generative AI and Competition,” marking a significant step in clarifying the agency’s stance on emerging AI-related antitrust issues. This commentary summarizes key takeaways of the report and discusses some points for further consideration. The takeaways are distilled into six key points as follows: (i) vigilance over NVIDIA's dominance in the AI chip market; (ii) caution regarding AWS's cloud computing market leadership; (iii) concerns over data access and data exploitation; (iv) expertise poaching, expertise concentration, and 'acqui-hiring'; (v) issues in the foundation models market; and (vi) vigilance over vertical expansion by Big Tech. This report touches upon various globally discussed issues, drawing upon other agencies' study results. However, there are still some points not discussed in the paper, raising questions about their omission. Also, it is notable that the competition authority, KFTC, appears to walk a fine line between competition policy and protectionism throughout the document. Continued updates and further consideration of AI competition policy will likely be necessary.	2025	Sangyun Lee	KFTC's Generative AI and Competition Report: A Quick Overview
Principles for AI Regulation and Competition Law AI-facilitated manipulation	Echo Chambers and Competition Law: Should Algorithmic Choices be Respected?	Algorithms are employed by a growing number of firms in order to make choices for users. One prominent example involves news and views consumption through media platforms, which is increasingly mediated by algorithmic personalization. Rather than engaging with the rich variety of ideas on the web, many online users are exposed primarily to content chosen by algorithms, which generally attempts to fit each user’s pre-existing views. This raises questions regarding competition law’s responsibility and ability to protect the free exchange of ideas in the marketplace. We argue that while competition law can be used to protect the diversity of content in the market, the protection of a diversity of exposure is much more challenging. An interference that is aimed at exposing users to diverse ideas need not conflict with the goals of antitrust. In particular, we argue that even if algorithmic choices attempt to cater to users’ preferences, they need not conflict with the ideals of consumer sovereignty, autonomy and choice on which competition law is based. Yet competition laws do not possesses the right tools to tackle this problem, and therefore doing so may best be left to other mechanisms. The discussion has implications for other algorithmic choices which may seem to cater to users’ preferences.	2020	Eran Fish, Michal S Gal	Echo Chambers and Competition Law: Should Algorithmic Choices be Respected?
Principles for AI Regulation and Competition Law	Comments on CMA Report on How Algorithms May Reduce Competition and Harm Customers	The Competition and Markets Authority (UK) has published their Research and analysis on ‘Algorithms: How they can reduce competition and harm consumers’. They have done so in parallel with a call for information on this critical area of research and potential regulation/standards. This article does two things: i. It provides a condensed summary of the report, and, ii. It notes our main findings and comments, which we offer as part of our feedback to the call for information. The principal takeaway is the need for disambiguation regarding whether or not novel and enforceable legislation/regulation/standards is in the pipeline and that empirical work in this area is greatly needed.	2021	Emre Kazim, Jeremy Barnett, Adriano Koshiyama	Comments on CMA Report on How Algorithms May Reduce Competition and Harm Customers
Principles for AI Regulation and Competition Law	Unfolding Digital Ignorance: How to Ensure the Accountability of Pricing Algorithms	By taking roles and responsibilities away from humans, self-learning pricing algorithms embed digital ignorance in pricing work. Digital ignorance may form a plausible defense for managers in deflecting blame and demonstrating their lack of awareness. While, in many cases, managers cannot know everything occurring in their organization concerning pricing, should cases of digital ignorance be allowed to flourish? If the answer to this question is no, how can the accountability of pricing algorithms be ensured? In analyzing these two research questions, this paper posits two contributions. First, it illustrates four aspects of digital ignorance induced by pricing algorithms. Second, the article proposes a sociotechnical processual approach to ensure the accountability of algorithms. Such a processual approach could dissuade managers and complicit parties from avoiding and deflecting blame.	2024	Vikash Kumar Sinha, Petri Kuoppamäki	Unfolding Digital Ignorance: How to Ensure the Accountability of Pricing Algorithms
Principles for AI Regulation and Competition Law AI and market power	Algorithms and Artificial Intelligence: An Optimist Approach to Efficiencies	Artificial intelligence (“Al") is a comprehensive field of computer science that enables machines to learn from experience and ultimately take difficult actions without human intervention. Al has been one of the highly popular topics over the last decade, as it has become a more effective tool after the development of algorithms (i.e.the bricks of an Al system) that teach machines to learn. That, in turn, has led a subfield of AI “machine- learning” and even “deep learning” (a subfield of machine learning) to arise, enabling machines to envisage human preferences and accordingly make decisions often more rapidly and efficiently than humans. Algorithms have been designed to carry out complex tasks and process data that can eventually reduce costs and bring certain advantages for both businesses and consumers. Therefore, Al and intelligent algorithms have been evaluated by many to have a significant power to realize policy goals of a society. Although many regulators and scholars approach the rapid development of algorithms and particularly Al with a bit of skepticism, focusing on their potential collusive effect in particular, this article aims to explore the pro-competitive impact of algorithms and Al and the efficiencies to be gained through the use of these technologies	2019	Gonenc Gurkaynak	Algorithms and Artificial Intelligence: An Optimist Approach to Efficiencies
Principles for AI Regulation and Competition Law	The Future of Cartel Deterrence and Detection	Over the last few decades, leniency programs recorded a successful history of identifying and dismantling cartels. The essential idea is simple, that authorities will reward a cartel member who self-reports. It is instructive to consider why authorities have relied so heavily on leniency in the past. First, it is not particularly resource intensive to implement. It doesn’t require collecting large amounts of data and employing economists and data analysts to sift through the haystack in hopes of finding the occasional needle. Second, almost by definition leniency is likely to have a high success rate of prosecution, among those applications selected to be fully investigated. It is noteworthy though that authorities are reluctant to produce statistics on the overall efficacy of their leniency programs. We should have a better idea of how many leniency applications are reviewed and investigated (among all those filed), and of those, how many lead nowhere. But we are not privy to such valuable information. Can authorities continue to almost exclusively depend on leniency programs going forward? There is no other area of criminal investigation which essentially waits for the guilty to confess as its key detection tool, yet the advantages of resource and success of such “passive detection” would equally apply to robbery or homicide: it doesn’t cost much to wait for a confession, and if someone confesses, the case will almost certainly be closed successfully. But while the police surveil neighborhoods to monitor possible illegal conduct, ready to not only detect ongoing conduct but also hopefully to deter such conduct from even getting started, several competition authorities still tend to lack the proactive nature of detection and deterrence through screening or market monitoring, relying (almost or entirely) on leniency programs. Hence, cartel detection appears, at least at first, to be a uniquely passive area of law enforcement. In this short paper we explore the role of leniency programs in the next generation of cartel detection, and ultimately deterrence. Will it continue to be the dominant source of cartel detection, or will advances in data collection and analysis – so-called “big data” and “machine learning” – reduce the cost and increase the effectiveness of screening and artificial intelligence techniques? Will traditional leniency and whistleblower programs even remain effective in a future which may keep no “paper trail” of communications proving the necessary intent? Have we learned the lessons from extensive rigging in financial markets? Do we need to revisit the entire detection approach?	2019	Rosa M Abrantes-Metz, Albert Metz	The Future of Cartel Deterrence and Detection
Principles for AI and Competition Law Governance	Antitrust by Algorithm	Technological innovation is changing private markets around the world. New advances in digital technology have created new opportunities for subtle and evasive forms of anticompetitive behavior by private firms. But some of these same technological advances could also help antitrust regulators improve their performance in detecting and responding to unlawful private conduct. We foresee that the growing digital complexity of the marketplace will necessitate that antitrust authorities increasingly rely on machine-learning algorithms to oversee market behavior. In making this transition, authorities will need to meet several key institutional challenges—building organizational capacity, avoiding legal pitfalls, and establishing public trust—to ensure successful implementation of antitrust by algorithm	2022	Cary Coglianese, Alicia Lai	Antitrust by Algorithm
Principles for AI and Competition Law Governance AI facilitated collusion	Algorithmic Competition	Companies increasingly use algorithms to set prices and create or enhance new products and services. While algorithms can result in many efficiency-enhancing and pro-competitive effects, they can also be used by firms to restrict competition. Competition authorities should be aware of these risks, know how to investigate them and identify any possible harm to consumers, as well as consider how to address this harm. This background note considers these important issues. It was prepared for discussions on “Algorithmic Competition” taking place at the June 2023 OECD Competition Committee.	2023	OECD	Algorithmic Competition
Principles for AI and Competition Law Governance AI and market power	Algorithms and Competition in the Digital Economy	The global economy is increasingly a digital economy driven by algorithms. This shift to a digital or algorithmic economy poses some distinct implications for how antitrust and consumer protection law evolves in the future. With this Foreword to a special issue published by Concurrences , we highlight major antitrust-related legal developments occurring around the world in response to the rapidly emerging environment of algorithmic-driven commerce. Without necessarily endorsing nor rejecting any of the various policies or proposals that have occurred in recent years, we organize and describe key antitrust-related legal developments that have arisen in response to the growth of the digital economy. In Part I, we detail some of the major legal changes or proposed changes that have targeted digital technology firms. Although many of these targeted firms deploy services that use algorithmic tools, competition authorities have not yet begun to do as much to regulate algorithmic services themselves as to target the firms that make use of them. And even though the specifics of some of the regulatory actions targeting digital firms can be said to be distinctive in their focus on online and other digital businesses, many of the concerns underlying regulatory actions or proposals have been, to date, similar to those that have long applied to general business activity. In Part II, we highlight an aspect of antitrust that might become truly novel in an increasingly algorithmic economy: the targeting of antitrust law and principles to business actions driven by algorithms themselves. As algorithmic tools come to automate economic transactions and autonomously make business decisions, the object of governmental oversight may well shift from the traditional focus on human managers to machine ones—or perhaps to the human designers of machine-learning “managers.” This is an emerging possibility which to date can be most saliently seen in the context of self-preferencing algorithms. Although antitrust enforcers appear thus far to target types of self-preferencing behaviors that have emanated from human decisions rather than fully independent algorithmic ones, it is not hard to conceive a future in which AI autonomously drives business decisions in problematic, anticompetitive directions or that operate on their own to charge supracompetitive prices. Finally, in Part III, in the face of a future that seems likely to be dominated by algorithmic transformations throughout the economy, antitrust regulators can expect to face a growing need themselves to develop and rely upon artificial intelligence and other algorithmic tools. The transition to an algorithmic economy, in the end, not only raises new sources of concern about competition and consumer protection, but it may also provide government with new opportunities to use digital tools to advance the goals of fair and efficient economic competition.	2024	Cary Coglianese, Alicia Lai	Algorithms and Competition in the Digital Economy
Principles for AI and Competition Law Governance AI and market power	Competition policy in a world of big data	Big data is an important phenomenon injecting transformative effects into social and economic relationships. Consumers, firms and machines produce unprecedented amounts of data collected, stored and analysed by leveraging the synergic capabilities of mathematics, computer science and the Internet. With the full advent of the Internet of Things, even more data will be observed about, and inferred from, individuals’ everyday activities and habits. The implied promise of big data is that it is increasingly possible to gain valuable insights out of unstructured data collected from different sources. Firms in many industries are increasingly using computer algorithms and big quantities of data to handle problems of analysis and prediction, from market intelligence to strategic management and automated decision-making. Acknowledging the growing potential for big data to have an immediate and direct impact on a broad range of human interactions, conversations within policy circles are starting to focus on how this phenomenon should factor into the competition policy framework itself. While big data can enhance competition, improve product offerings, and create a marketplace where resources are allocated more efficiently, the Chapter argues that competition policy designers and enforcers are bound to deal with unprecedented data-related challenges. The Chapter starts with a description of the big data value chain, highlights in particular how data collection, storage and analysis are driving many of the multisided business models of the digital economy, summarises some well-known peculiarities of data as an economic asset and sets the framework for the analysis of the effects of big data on competition processes. The Chapter concludes by drawing a few preliminary implications for competition policy. In particular, big data could have the effect of making collusion more prevalent, stable and difficult to detect, of reshaping traditional relationships within a vertical supply chain by increasing forms of dependency and potentially restraining inter-platform competition and users behaviour, of increasing market concentration, and, finally, of enabling further abuses of market power.	2016	Simonetta Vezzoso	Competition policy in a world of big data
Principles for AI and Competition Law Governance AI and market power	Competition by Design	Algorithms are the fundamental ingredient of online businesses such as search engines, marketplaces, peer-to peer platforms and social networks. They have already deeply affected the way individuals shop, communicate, and interact with one another. In pursuit of automation-driven efficiencies and market opportunities, firms are increasingly implementing algorithmic solutions that allow them to engage more effectively in commercial activity. In particular, algorithms can be an important source of innovation, allowing companies to develop non-traditional business models and extract more value from data, in order to improve product quality and customisation. Admittedly, advanced algorithms can be powerful tools in the hands of market players, and the challenges posed by algorithms are not entirely new problems for competition enforcers. Whereas the issues of algorithmic transparency and accountability are common to other areas of law and policy, there are further and more specific implications for competition policy. Price-setting algorithms in particular have drawn the attention of academics and policy makers. The current discussions and in-depth reflections, however, are not exclusively on the topic of pricing algorithm and collusion, but include a broader range of questions that are possibly going to inform effective competition policy in the age of algorithms. Against the background of increasingly algorithm-based markets, the paper aims to explore the potential for effective competition policy of the notion of “competition by design”, as broadly derived from the related concept of “data protection by design,” which is enshrined in the upcoming General Data Protection Regulation. EU Competition Commissioner Margarethe Vestager made clear in a series of recent public interventions that firms applying algorithms need to think from the start about how to keep them compliant with competition law (“algorithms will have to go to law school before they are let out”). While the idea of competition compliance by design might be gaining some foothold in the mind-sets of some competition authorities, there are currently no clear indications how it could be integrated into the already complex competition policy fabric. The paper concludes that algorithm design thinking could be a promising new tool at the disposal of competition authorities in the digital economy.	2017	Simonetta Vezzoso	Competition by Design
Principles for AI and Competition Law Governance	OECD Competition Trends 2025	This 2025 edition of the OECD Competition Trends report highlights worldwide competition enforcement trends during the calendar year 2023 based on 69 jurisdictions. Similar to previous editions, this year’s report identifies trends over time. Analyses focus on enforcement activity in cartels, abuse of dominance cases, mergers and advocacy activity. Moreover, this year’s edition includes a special chapter on the evolution of competition authorities’ resources, exploring potential reasons for the overall increasing trend for average budgets and staff dedicated to competition. The report contributes to continuously improving competition law and policy around the world.	2025	OECD	OECD Competition Trends 2025
Principles for AI Regulation and Competition Law	Applying EU Competition Law to Online Platforms: The Road Ahead	The adequate application of EU competition law to online platforms will entail multiple challenges to the current framework, which have yet been comprehensively explored. This contribution provides an overview of such challenges that enforcement authorities and undertakings will face in absence of certain modifications to the framework of Articles 101 and 102 TFEU. The various challenges that will arise in the course of application may be divided into three categories, namely application thresholds, qualification of practices and justification possibilities. In the context of this article, application thresholds refer to the instance wherein certain behavior is considered as falling under the scope of Art. 101 or 102 TFEU. Qualifications of practices refer to the assessment of an investigated behavior and the finding of an abuse of dominance or a restriction of competition through coordination by object or effect. Finally, challenges concerning justification grounds in the context of this paper refer to the feasibility of relying on the justification grounds of Articles 101 and 102 TFEU.	2018	Daniel Mandrescu	Applying EU Competition Law to Online Platforms: The Road Ahead
Principles for AI Regulation and Competition Law	Algorithmic Consumers	The next generation of e-commerce will be conducted by digital agents, based on algorithms that will not only make purchase recommendations, but will also predict what we want, make purchase decisions, negotiate and execute the transaction for the consumers, and even automatically form coalitions of buyers to enjoy better terms, thereby replacing human decision-making. Algorithmic consumers have the potential to change dramatically the way we conduct business, raising new conceptual and regulatory challenges. This game-changing technological development has significant implications for regulation, which should be adjusted to a reality of consumers making their purchase decisions via algorithms. Despite this challenge, scholarship addressing commercial algorithms focused primarily on the use of algorithms by suppliers. This article seeks to fill this void. We first explore the technological advances which are shaping algorithmic consumers, and analyze how these advances affect the competitive dynamic in the market. Then we analyze the implications of such technological advances on regulation, identifying three main challenges.	2020	Michal Gal, Niva Elkin-Koren	Algorithmic Consumers
Principles for AI Regulation and Competition Law AI and market power	Challenges for Competition Law Enforcement and Policy in the Digital Economy	Digitalisation, new technologies and scientific breakthroughs are unfolding on many fronts. Advances in communication and data processing are not just profoundly affecting existing industries, but also rearranging global value chains, thereby allowing for entirely new products and services and disrupting traditional ones. These trends can deliver benefits and stimulate economic growth; however, they can also give rise to competition concerns as well as create needs for new regulation. Since the effects on society extend far beyond the digital technology context alone, concerns arising from digitalisation have become increasingly relevant for both policy makers and stakeholders. The OECD has recognised in multiple occasions the impact of digitalisation and data-driven innovations on competition. Recent policy discussions have covered a wide range of policy issues (e.g. disruptive innovation, two-sided markets, Big Data, algorithms and collusion, enforcement tools in multi-sided markets) as well as sector-specific topics (e.g. legal services, financial markets, road transport, and electricity). In 2016, the digital economy was selected as a strategic theme for the OECD Competition Committee focusing on four sub-streams: the relationship between the digital economy, competition law and innovation, challenges posed to prevailing antitrust tools and approaches, practical challenges to competition enforcement, and development and evolution of specific industries. This theme also fits in a broader OECD-wide Going Digital project, which aims at developing a comprehensive policy approach to the digital transformation and includes ‘deep dives’ into some of the challenges of the digital era, such as jobs and skills in the digital economy and the implications of the digital transformation for competition and market openness	2017	Antonio Capobianco, Anita Nyeso	Challenges for Competition Law Enforcement and Policy in the Digital Economy
Principles for AI Regulation and Competition Law AI and market power	The Generative AI challenges for competition authorities	Generative Artificial intelligence (GenAI) stimulates market and regulatory developments. It has spurred competition and innovation and created new emerging markets and technologies ranging from chips, machine learning models, cloud computing, and software to answer engines. Market developments pose competition issues. Developers need access to three key components in the value chain: computing resources, machine learning models, and data. Despite various players in the field, certain sectors face competition concerns due to market features and potential issues, such as tying. Regulatory developments impact GenAI. In particular, GenAI raises unsolved regulatory challenges arising from the use of copyrighted data (intellectual property rights), personal data (data protection), AI risks (AI governance) and competition. Competition authorities closely monitor GenAI developments by doing market studies. These initiatives will inform how competition in GenAI works. Challenges for competition authorities stem from emerging markets and technologies, along with regulatory instabilities. New products and services are shaping new and existing markets, like answer engines and advertising. Regulatory uncertainties influence competition in GenAI. At this current development stage, competition authorities should focus on understanding market and regulatory developments by cooperating among themselves and with relevant competent authorities. They should exercise formal enforcement powers and potentially update competition tools only when necessary and justified, guided by the insights gained from these market studies.	2024	Christophe Carugati	The Generative AI challenges for competition authorities
Principles for AI Regulation and Competition Law AI and market power	Imposing Regulation on Advanced Algorithms	This book discusses the necessity and perhaps urgency for the regulation of algorithms on which new technologies rely; technologies that have the potential to re-shape human societies. From commerce and farming to medical care and education, it is difficult to find any aspect of our lives that will not be affected by these emerging technologies. At the same time, artificial intelligence, deep learning, machine learning, cognitive computing, blockchain, virtual reality and augmented reality, belong to the fields most likely to affect law and, in particular, administrative law. The book examines universally applicable patterns in administrative decisions and judicial rulings. First, similarities and divergence in behavior among the different cases are identified by analyzing parameters ranging from geographical location and administrative decisions to judicial reasoning and legal basis. As it turns out, in several of the cases presented, sources of general law, such as competition or labor law, are invoked as a legal basis, due to the lack of current specialized legislation. This book also investigates the role and significance of national and indeed supranational regulatory bodies for advanced algorithms and considers ENISA, an EU agency that focuses on network and information security, as an interesting candidate for a European regulator of advanced algorithms. Lastly, it discusses the involvement of representative institutions in algorithmic regulation.	2019	Fotios Fitsilis	Imposing Regulation on Advanced Algorithms
Principles for AI Regulation and Competition Law AI facilitated collusion	Competition Law Consequences of Artificial Intelligence	Technology, whose etymology derives from Aristotle’s “techne,” (which means “craftsmanship” or “art”) represents one of the most primal instincts and distinctive features of human beings, which is the ability to use tools to increase one’s quality of life. In fact, today we can observe the astonishing growth of the digitalized economy, and the emergence (and rapid adoption) of various novel commercial activities that are enabled by new technologies, such as data collection or multi-sided markets, which have resulted in significant enhancements in total welfare. Following the emergence and rapid development of computer sciences, artificial intelligence has become an increasingly important part of our daily lives. Algorithms will change the competition landscape as we know it by enabling undertakings to improve their pricing models, by helping suppliers to provide better products and services to their consumers, and by predicting market trends to encourage and spur innovation.	2018	Gonenc Gurkaynak, Iskin Naz Altinsoy, Umay Rona	Competition Law Consequences of Artificial Intelligence
Principles for AI Regulation and Competition Law AI and market power	Comparative Analysis Regulatory of AI and Algorithm in UK, EU and USA	Regulation may mean anything from setting limits on what can be created to determining which tasks must remain in the hands of humans. Lessig proposes the Techno regulatory approach, which protects the AI through making policies that affect the code with which the AI is programmed. Governments and businesses use algorithms to make decisions that substantially impact our lives. There is currently no regulation to regulate the use of artificial intelligence. On the other hand, AI systems are governed by current laws - data protection, consumer protection, and market competition rules. This study research the Algorithmic decision-making uses algorithms to execute or inform judgments that might range from simple to complex. Tim O'Reilly urges the government to look into algorithm regulation, but he never defines the term. What structure should we use when dealing with circumstances where governmental and non-governmental authorities remain intimately connected? The UK government has released its National AI Strategy, which outlines its plan to transform the UK into a global Artificial Intelligence powerhouse over the next ten years. There are no particular regulations on the regulation of AI and the algorithms; however, soft law, ethical conduct, and platform regulations exist. The EU proposal is for AI regulation that is cross-sectoral and risk-based. Critical hazards associated with AI use and liability-related issues apply standards that protect fundamental rights. The act guarantees that AI is legal, ethical, and technically sound while also upholding democratic principles, human rights, and the Rule of Law. The European Artificial Intelligence Board (EAIB) would be established as a new enforcement authority at the Union level. National supervisors will flank EAIB at the Member State level. Fines of up to '6% of global turnover, or 30 million euros for individual corporations' can be imposed.	2022	Olanrewaju Akinola, Ogundipe Adebayo Tunbosun, Bankole Oladapo	Comparative Analysis Regulatory of AI and Algorithm in UK, EU and USA
Principles for AI Regulation and Competition Law AI and market power	Framing Algorithms – Competition Law and (Other) Regulatory Tools	As other fields of law, competition law is put to the test by new technologies in general and algorithmic market activity in particular. This paper takes a holistic approach by looking at areas of law, namely financial regulation and data protection, which have already put in place rules and procedures to deal with issues arising from algorithms. Before making the bridge and assessing whether the application of any such tool might be fruitful for competition law, the paper discusses important competition cases regarding algorithms, including the Google Shopping, Lufthansa and Facebook case. It concludes with some policy recommendations.	2018	Peter Georg Picht, Gaspare Loderer	Framing Algorithms – Competition Law and (Other) Regulatory Tools
Principles for AI Regulation and Competition Law AI and market power	Antitrust and Artificial Intelligence - A Research Agenda	This short paper provides an overview of the claims made in the emerging scholarship on artificial intelligence and antitrust policy. It discusses in particular the conjecture that markets will be rife with algorithmic collusion and extractive personalized pricing. It stresses the main methodological, theoretical and empirical limitations that underpin the claims made in the scholarship, and concludes with a call for evidenced-based antitrust policy.	2020	Nicolas Petit	Antitrust and Artificial Intelligence - A Research Agenda
Principles for AI Regulation and Competition Law AI and market power	It's a Feature, not a Bug: On Learning Algorithms and what they teach us	Algorithms, the recipe of any computational program, are considered to be an essential component of computer science . Algorithms may be simple, such as those for sorting integers, or highly complex. What they all share is an exact sequential set of commands that allow a computer system to perform computational tasks. The efficiency of an algorithm is affected, inter alia, by the availability of data structures, which allow storing and retrieving data efficiently, and by the level of its complexity, which determines the time it may take an algorithm to perform its task in a worst-case scenario (which may be forever). Algorithms come in various shapes and forms. Some algorithms are monolithic, encompassing every aspect of the computational process. Others may perform only part thereof, leaving other tasks to other algorithms. Some are designed for sequential execution while others are meant to be run distributed and in parallel. Yet, their main essence remains the same: given an appropriately designed input, a sequence of commands is performed over this input to generate an output in a clearly defined format. Recent years have seen an increased interest in a specific class of algorithms, Machine Learning (ML) algorithms . Such algorithms are at the backend of what many believed to be part of human beings realm alone. IBM introduced Watson, a winner of the American popular show Jeopardy. Google introduced AlphaGo, a Go winner . Autonomous driving vehicles are no longer science fiction, language translators have shown an immense improvement, and we all have our personal assistants listening and reacting to our vocal commands over the phone. The impact ML algorithms have on the new digital economy stretches far beyond what many sceptical researchers thought possible. So much so that ML algorithms have invoked many primeval concerns and fears brought forward in well-known movies such as the Matrix. The purpose of this short note is to enhance the understanding of what it means for an algorithm to learn, who teaches the algorithms, what they learn, and to what extent they can share with us what they have learned. This understanding can serve as a starting point for regulators to better understand what can and cannot be demanded from market players using algorithms. The discussion focuses in particular on machine learning algorithms. The reason is three-fold. First, the use of such algorithms is becoming more commonplace. Second, they often obscure the reasons for their outcomes, making it easier for the user to claim he was not aware of what the outcome would be. Finally, such algorithms can be employed by regulatory authorities to study other algorithms or to determine what has driven the market outcome.	2017	OECD Avigdor Gal	"It's a Feature, not a Bug: On Learning Algorithms and what they teach us
Principles for AI Regulation and Competition Law	Artificial Intelligence and Competition Law	Firms use algorithms for important decisions in areas from pricing strategy to product design. Increased price transparency and availability of personal data, combined with ever more sophisticated machine learning algorithms, has turbocharged their use. Algorithms can be a procompetitive force, such as when used to undercut competitors or to improve recommendations. But algorithms can also distort competition, as when firms use them to collude or to exclude competitors. EU competition law, in particular its provisions on restrictive agreements and abuse of dominance (Articles 101–102 TFEU), prohibits such practices, but novel anticompetitive practices—when algorithms collude autonomously for example—may escape its grasp. This chapter assesses to what extent anticompetitive algorithmic practices are covered by EU competition law, examining horizontal agreements (collusion), vertical agreements (resale price maintenance), exclusionary conduct (ranking), and exploitative conduct (personalized pricing).	2025	Friso Bostoen	Artificial Intelligence and Competition Law
Principles for AI Regulation and Competition Law AI and market power	Decoding the AI Act: Implications for Competition Law and Market Dynamics	The AI Act is poised to become a pillar of modern competition law. The present article seeks to offer those interested in competition law with a critical guide to its key provisions. It also discusses the AI Act’s effect on innovation and competition within the EU single market.	2025	Thibault Schrepel	Decoding the AI Act: Implications for Competition Law and Market Dynamics
Principles for AI Regulation and Competition Law	Die Entwicklung des europäischen Kartellrechts	As a follow-up to EuZW 2023, 355, this article summarizes the most relevant developments of European competition law (including DMA) during the year 2023. As usual, the areas of EU merger control and EU State aid are not covered (regarding state aid see most recently Soltész in EuZW 2024, 5).	2016	Andreas Weitbrecht,Jan Mühle	Die Entwicklung des europäischen Kartellrechts

AI and remedies

Category of Compe-titive concern

Link

Abstract

Year

Author or Insitutution

Title

AI and remedies

AI facilitated collusion

A Critique of Recent Remedies for Third-Party Pricing Algorithms and Why the Solution is not Restrictions on Data Sharing

A growing antitrust challenge is competitors using a pricing algorithm supplied by the same data analytics company. While there can be procompetitive efficiencies in outsourcing pricing, the risk of anticompetitive harm in having a common agent influence competitors' prices is severe. To deal with this challenge, a remedy has recently been proposed in the United States at the federal level and is being adopted at the local level. This remedy prohibits a third party's use of nonpublic competitor data. If firms A and B both subscribe to the same third party, that third party cannot use the nonpublic data of firm B in the pricing algorithm it uses to recommend prices to firm A. The contribution of this paper is to critically examine this remedy. First, it is explained that the remedy creates inefficiencies which need to be recognized. Second, and more importantly, it is shown the remedy may not prevent the harm it is intended to prevent. More specifically, a workaround is developed whereby a third party can result in firms charging supracompetitive prices while not using nonpublic competitor data and thus abiding by the law. The problem is that the remedy focuses on shared data when the source of harm is shared objective. A legal scholar and I are in the process of developing a remedy based on prohibiting shared objective which will be described in a future paper.

2025

Joseph E Harrington Jr

A Critique of Recent Remedies for Third-Party Pricing Algorithms and Why the Solution is not Restrictions on Data Sharing

Other

Category of Compe-titive concern	Link	Abstract	Year	Author or Insitutution	Title
-	Deterministic limit of temporal difference reinforcement learning for stochastic games	Reinforcement learning in multiagent systems has been studied in the fields of economic game theory, artificial intelligence, and statistical physics by developing an analytical understanding of the learning dynamics (often in relation to the replicator dynamics of evolutionary game theory). However, the majority of these analytical studies focuses on repeated normal form games, which only have a single environmental state. Environmental dynamics, i.e., changes in the state of an environment affecting the agents' payoffs has received less attention, lacking a universal method to obtain deterministic equations from established multistate reinforcement learning algorithms. In this work we present a methodological extension, separating the interaction from the adaptation timescale, to derive the deterministic limit of a general class of reinforcement learning algorithms, called temporal difference learning. This form of learning is equipped to function in more realistic multistate environments by using the estimated value of future environmental states to adapt the agent's behavior. We demonstrate the potential of our method with the three well-established learning algorithms Q learning, SARSA learning, and actor-critic learning. Illustrations of their dynamics on two multiagent, multistate environments reveal a wide range of different dynamical regimes, such as convergence to fixed points, limit cycles, and even deterministic chaos.	2019	Wolfram Barfuss, Jonathan F. Donges, Jürgen Kurths	Deterministic limit of temporal difference reinforcement learning for stochastic games
-	Prediction of Fuel Efficiency Using Machine Learning Algorithms	The auto industry is highly volatile. Automobile manufacturers are continually improving their procedures to improve fuel efficiency in response to rising gasoline prices and to stabilize consumer churn. What if, however, one might have identified a trustworthy approximator for a car's miles per gallon given some well-known details about the car? Then, one may outperform the competition through offering a more appealing and effective vehicle, drastically cutting excessively wasteful research & development costs, and most significantly, seizing a sizable portion of the market. Machine learning has long been integrated into the automotive business, and it has been shown to be successful in every way. The problems faced in the automobile industry have been scaled down and fed to an Applied Machine Learning approach for prediction of fuel efficiency by optimal use of available resources. The UCI Machine Learning Repository: Auto MPG dataset has been selected as the test subject. The following criteria were met during the process: (a) Determine which Machine Learning Model is most optimal. (b) Observe how data collection affects the outcome. (c) Evaluate salient features. The Multivariate dataset comprises 3 multi-valued discrete attributes and 5 continuous attributes. A couple attributes were extracted from the raw dataset. Acceleration on power (an attribute mined from the raw dataset) proved to be the best estimator among the attributes. Regression is applied on the dataset, where models take turns to predict fuel consumption accurately. The absolute relative error of different machine learning algorithms is found to be less than 5%. The random forest model is appropriate for widespread application because of its high score in accuracy, sensitivity and specificity. The machine learning framework that has been implemented sets the groundwork for improving the monitoring of automobile databases and precise management of fuel consumption in urban and rural transportation systems.	2023	Krishnanunni R.Nair, Manbir Kaur	Prediction of Fuel Efficiency Using Machine Learning Algorithms
-	Here’s why algorithms are not (really) a thing	-	2017	Thibault Schrepel	Here’s why algorithms are not (really) a thing’

AI Database

On this page